defining key molecules in a myeloma cell nicheetheses.whiterose.ac.uk/6732/1/julia phd thesis...

Defining Key Molecules in a

Myeloma Cell Niche

Julia Paton-Hough

The Department of Human Metabolism

Thesis submitted to The University of Sheffield for the degree

of Doctor of Philosophy

August 2014

Presentations and publications Page i

Presentations and publications

Presentations:

1. J. Paton-Hough, C. A. Eaton, P.I. Croucher and M. A. Lawson. Identification of

potential candidate molecules associated with myeloma cell engagement in a

bone marrow niche. Herbert Fleisch Workshop, Bruges, Belgium, 2014. Oral and

poster presentation.

2. J.Hough, M.A.Lawson, C.A. Eaton & P. Croucher. Identification of potential

candidate molecules associated with myeloma cell engagement in a bone

marrow niche. Joint meeting of the Bone Research Society and the British

Orthopaedic Research Society, Oxford, UK 2013. Poster presentation.

3. J. Hough, C. A. Eaton, P.I. Croucher and M. A. Lawson. Identification of potential

candidate molecules associated with myeloma cell engagement in a bone

marrow niche. International conference of Cancer-induced bone disease,

Miami, Florida, USA, 2013. Poster presentation.

4. J. Hough, O.Al-Amer M.A. Lawson, C.A.Eaton and P.I. Croucher. Identification of

molecular determinants of myeloma cell engagement in the osteoblast niche in

bone. Mellanby Centre Annual Research Day, Sheffield, UK, 2012. Poster

presentation.



bone. Annual School Research Day, Sheffield, UK, 2012. Poster presentation.



Presentations and publications Page ii

bone. Mellanby Centre Annual Research Day, Sheffield, UK, 2012. Oral and




bone. Mellanby Centre Annual Research Day, Sheffield, UK, 2011. Oral and




bone. Annual School Research Day, Sheffield, UK, 2011. Poster presentation.

Publications:

1. NOD/SCID-GAMMA mice are an ideal strain to assess the efficacy of therapeutic

agents used in the treatment of myeloma bone disease. M. A. Lawson, J. Paton-

Hough, H. Evans, R.E. Walker, W. Harris, J. A. Snowden & A.D. Chantry. Submitted

and under review at Plos One.

2. A Novel Method for Longitudinal Tracking of Dormant Cancer Cell Dynamics in

the Skeleton as a Platform for Drug Discovery. Michelle A. Lawson, Michelle M.

McDonald, Natasa Kovacic, Weng-Hua Khoo, Jenny Down, Warren Kaplan, Julia

Paton-Hough, Clair Fellows, T. Neil Dear, Els Van Valckenborgh, Paul A. Baldock,

Mike J. Rogers, Colby L. Eaton, Karin Vanderkerken, Tri Giang Phan, Peter I.

Croucher. To be submitted to Cancer Discovery or similar.

3. A review of current murine models of myeloma used to assess the efficacy of

therapeutic agents on tumour growth and bone disease. J. Paton-Hough, A.D.

Chantry & M.A. Lawson: Invited review from The Journal of Laboratory

Haematology, to be submitted, August 2014.

Presentations and publications Page iii

4. N-cadherin is a key molecule for facilitating myeloma cell attachment to the

osteoblastic niche and subsequent tumour development in bone. J.Paton-

Hough, O.Al-Amer, P.I. Croucher, C.A.Eaton, M.A.Lawson. Currently being

written, to be submitted to Plos One or similar.

5. Myeloma cells express haemopoeitic stem cell markers of quiescence in vitro

and in vivo. J.Paton-Hough, P.I. Croucher, C.A.Eaton, M.A.Lawson. Currently

being written, to be submitted to Leukaemia Research or similar.

Acknowledgements Page iv

Acknowledgements

I would like to firstly thank Prof. Peter Croucher for supporting my PhD application,

which enabled me to conduct my PhD and to continue my technical role in the Bone

Analysis Laboratory, and for this, I am extremely grateful. I would also like to thank

him for his supervision and continual input despite leaving the University for warmer

climates.

Thanks to Dr Colby Eaton who became my primary supervisor, replacing Peter in my

second year. I am extremely grateful for all of the support and guidance he provided

as well as boosting moral with anecdotes about pork pies and holidays.

I would especially like to thank Dr Shelly Lawson for her continued guidance and

relentless support over the course of my PhD. I am particularly grateful for all of the

help she provided (especially in the marathon field lab sessions) as well as her

positive outlook in any situation, which helped me through some of the trying times

in my PhD.

I would like to acknowledge everyone in the newly established SMaRT group; which

includes Dr Andy Chantry, the usual source of entertainment (and by this I do not

mean his punk band), Dr Clive Buckle, Holly Evans and Darren Lath, particularly for

maintaining my sanity over the course of my PhD. I would also like to thank Orla

Gallagher in the Bone Analysis Laboratory for her support, in addition to her histology

training.

In addition, I would like to thank everybody in the Department of Human Metabolism

for all of the help provided as well as making the department a pleasant environment

to work in. This particularly includes Dr Gareth Richards for providing his expertise in

molecular biology and Dr Suruchi Pacharne for all of her help with Western Blotting.

I would also like to thank people who are sorely missed at the University, which

includes Dr Marco Baud’huin and Jay Gurubalan, for all of their help and guidance

while they were here.

Acknowledgements Page v

Finally, I would like to thank my (now) husband Tom Paton for all of his kindness,

compassion, endless support, and motivational speeches which I could not have

done without. I would also like to acknowledge my parents and new in-laws for their

constant support and kindness.

Apologies if there is anyone that I have forgotten! Thank you to everyone that helped,

guided, supported and offered a kind word to me over the course of my PhD.

Table of contents Page vi

Table of contents

Presentations and publications ..............................................................................

Acknowledgements ............................................................................................ iv

Table of contents ................................................................................................ vi

Abstract ............................................................................................................. xxi

Abbreviations .................................................................................................. xxiii

Chapter I: Introduction ........................................................................................ 1

1.1 Multiple Myeloma ...................................................................................... 2

1.1.1 Epidemiology of multiple myeloma ............................................................ 2

1.1.2 Pathology of disease ................................................................................... 2

1.1.3 Clinical features ........................................................................................... 4

1.1.4 Diagnosis ..................................................................................................... 5

1.1.5 Pathology of tumour-induced lytic bone disease ....................................... 7

1.1.6 Treatment options .................................................................................... 10

1.1.7 Treatments to reduce tumour burden ...................................................... 10

1.1.8 Treatments to alleviate symptoms ........................................................... 11

1.1.9 Recurrence of disease ............................................................................... 12

1.2 The HSC Niche .......................................................................................... 13

1.2.1 Evidence for a HSC niche ........................................................................... 13

1.2.2 The osteoblastic HSC niche ....................................................................... 15

Table of contents Page vii

1.2.3 The vascular HSC niche ............................................................................. 17

1.3 Key molecules in the osteoblastic niche .................................................... 18

1.3.1 CXCR4/CXCL12 ........................................................................................... 18

1.3.2 Notch-1/Jag-1 ............................................................................................ 21

1.3.3 Tie-2/Ang-1................................................................................................ 25

1.3.4 N-cad ......................................................................................................... 27

1.4 A potential myeloma cell niche ................................................................. 30

1.5 Preclinical multiple myeloma models ........................................................ 33

1.6 Summary .................................................................................................. 34

1.7 Aims, hypothesis and objectives ............................................................... 35

1.7.1 Aims and hypothesis ................................................................................. 35

1.7.2 Objectives .................................................................................................. 35

Chapter II: Materials and methods ..................................................................... 37

2.1 General cell culture methods .................................................................... 38

2.1.1 Cells ........................................................................................................... 38

2.1.1.1 The 5T murine myeloma cell lines ..................................................... 38

2.1.1.2 MC3T3-E1 osteoblast cell line ............................................................ 38

2.1.1.3 Primary osteoblast lineage cells......................................................... 38

2.1.2 Cell culture ................................................................................................ 39

2.1.2.1 5T murine myeloma cell line culture ................................................. 39

Table of contents Page viii

2.1.2.2 MC3T3-E1 cell line culture ................................................................. 39

2.1.2.3 Primary osteoblast lineage cell culture .............................................. 40

2.1.2.4 Cell counting ....................................................................................... 40

2.1.2.5 Cell thawing ........................................................................................ 40

2.1.2.6 Cell freezing ........................................................................................ 41

2.1.2.7 Primary osteoblast lineage cell differentiation .................................. 41

2.2 Primary animal cells and tissues ............................................................... 41

2.2.1 Animal cell isolation .................................................................................. 41

2.2.1.1 Animals ............................................................................................... 42

2.2.1.2 Primary osteoblast lineage cell isolation ........................................... 42

2.2.1.3 Bone marrow isolation for RNA extraction ........................................ 43

2.2.1.4 Bone marrow, splenocyte and blood isolation for flow cytometry ... 43

2.2.2 In vivo models of multiple myeloma ......................................................... 44

2.2.2.1 The 5T33-GFP and 5TGM1-GFP models of multiple myeloma for

detection of HSC niche molecules and ligands .............................................. 44

2.2.2.2 Gene knock-down experiments in the 5TGM1-GFP model ............... 45

2.3 Molecular biology methods ...................................................................... 45

2.3.1 RNA extraction .......................................................................................... 45

2.3.1.1 RNA methodology .............................................................................. 45

2.3.1.2 Quantification of RNA ........................................................................ 46

2.3.1.3 RNA fragment-length analysis ........................................................... 46

Table of contents Page ix

2.3.2 Reverse transcription ................................................................................ 47

2.3.2.1 Reverse transcription methodology .................................................. 47

2.3.3 Endpoint PCR ............................................................................................. 49

2.3.3.1 Principals of endpoint PCR ................................................................. 49

2.3.3.2 Endpoint PCR methodology ............................................................... 50

2.3.3.3 Primer design ..................................................................................... 51

2.3.4 Gel electrophoresis ................................................................................... 52

2.3.4.1 Principals of gel electrophoresis ........................................................ 52

2.3.4.2 Gel electrophoresis methodology ...................................................... 52

2.3.4.3 Gene sequencing ................................................................................ 53

2.3.5 Real-time PCR ............................................................................................ 53

2.3.5.1 Principals of real-time PCR ................................................................. 53

2.3.5.2 Real-time PCR methodology .............................................................. 54

2.3.5.3 Probe efficiency .................................................................................. 56

2.4 Cell biology .............................................................................................. 56

2.4.1 Flow cytometry ......................................................................................... 56

2.4.1.1 General principles of flow cytometry ................................................ 56

2.4.1.2 Flow cytometry methodology ............................................................ 58

2.4.2 Immunofluoresence staining .................................................................... 61

2.4.2.1 Principles of immunofluorescence staining ....................................... 61

Table of contents Page x

2.4.2.2 Immunofluorescence staining methodology ..................................... 61

2.4.2.3 Fluorescent microscopy ..................................................................... 63

2.4.3 IHC staining ............................................................................................... 65

2.4.3.1 Principles of IHC ................................................................................. 65

2.4.3.2 Preparation of bones for IHC ............................................................. 66

2.4.3.3 IHC methodology for CD138 .............................................................. 67

2.4.3.4 IHC methodology for N-cad ............................................................... 68

2.4.3.5 IHC methodology for CXCR4 .............................................................. 69

2.4.3.6 CD138 staining quantification ............................................................ 69

2.4.4 Western blotting ....................................................................................... 70

2.4.4.1 Principals of Western blotting ........................................................... 71

2.4.4.2 Cell lysis .............................................................................................. 71

2.4.4.3 Protein quantification ........................................................................ 72

2.4.4.4 Bicinchoninic acid assay methodology .............................................. 72

2.4.4.5 Acrylamide gel methodology ............................................................. 74

2.4.4.6 Sodium dodecyl sulphate-polyacrylamide gel electrophoresis ......... 75

2.4.4.7 Protein transfer .................................................................................. 76

2.4.4.8 Immuno-blot ...................................................................................... 77

2.4.4.9 Immuno-blot development ................................................................ 77

2.4.4.10 Western blotting quantification ...................................................... 78

Table of contents Page xi

2.4.5 Myeloma cell growth curves ..................................................................... 79

2.4.5.1 Growth curve methodology ............................................................... 79

2.5 Primary osteoblast lineage cell differentiation .......................................... 80

2.5.1 Alkaline phosphatase production analysis ................................................ 80

2.5.1.1 Principals of alkaline phosphatase analysis ....................................... 80

2.5.1.2 Alkaline phosphatase assay methodology ......................................... 81

2.5.1.3 DNA quantification methodology ...................................................... 82

2.5.1.4 Calculating alkaline phosphatase activity .......................................... 83

2.5.2 Primary osteoblast lineage cell mineralisation analysis ........................... 84

2.5.2.1 Primary osteoblast lineage cell mineralisation assay methodology .. 84

2.5.3 Myeloma and primary osteoblast lineage cell co-cultures ....................... 85

2.5.3.1 Myeloma and primary osteoblast lineage cell co-culture methodology

........................................................................................................................ 85

2.5.3.2 Co-culture analysis ............................................................................. 85

2.6 Knock-down of key HSC niche molecules .................................................. 88

2.6.1 Short hair-pin RNA gene knock-down technology .................................... 88

2.6.1.1 Principals of RNA interference technology ........................................ 88

2.6.2 Puromycin killing curves............................................................................ 90

2.6.2.1 Principals of puromycin selection ...................................................... 90

2.6.2.2 Puromycin killing curve methodology ............................................... 90

2.6.3. Polybrene optimisation ............................................................................ 90

Table of contents Page xii

2.6.3.1 Polybrene optimisation methodology ............................................... 91

2.6.4 Green fluorescent protein transduction in 5T33-WT cells and knock-down

of CXCR4 and N-cad in 5TGM1-GFP cells ........................................................... 91

2.6.4.1 Green fluorescent protein transduction and knock-down of CXCR4 and

N-cad in the 5TGM1-GFP cells methodology ................................................. 91

2.6.4.2 PCR analysis ........................................................................................ 93

2.7. Bone parameter analysis ......................................................................... 94

2.7.1 Micro-computed tomography .................................................................. 94

2.7.1.1 Principals of micro-computed tomography ....................................... 94

2.7.1.2 Micro-computed tomography methodology ..................................... 96

2.8. Statistics ................................................................................................. 97

Chapter III: The in vitro expression of HSC niche molecules and their

complementary ligands by myeloma and osteoblastic cells ................................ 98

3.1 Introduction ............................................................................................. 99

3.2 Hypothesis, aims and objectives ............................................................. 101

3.2.1. Hypothesis and aims .............................................................................. 101

3.2.2 Objectives ................................................................................................ 101

3.3 Methods ................................................................................................ 102

3.3.1 Gene expression analysis ........................................................................ 102

3.3.1.1 Endpoint PCR .................................................................................... 102

3.3.1.2 Real-time PCR ................................................................................... 102

Table of contents Page xiii

3.3.2 Protein quantification analysis ................................................................ 103

3.3.2.1 Flow cytometry ................................................................................ 103

3.3.2.2 Immunofluorescence ....................................................................... 103

3.3.3. Statistics ................................................................................................. 103

3.4 Results ................................................................................................... 104

3.4.1 Qualitative assessment of gene expression for the HSC niche molecules,

ligands and CD138 by normal murine constituents and murine multiple

myeloma cells, determined using endpoint PCR ............................................. 104

3.4.2 Quantitative assessment of gene expression for the HSC niche molecules,

ligands and CD138 by normal murine constituents and murine multiple

myeloma cells, determined using real-time PCR ............................................. 105

3.4.3 Protein quantification of HSC niche molecules, ligands and CD138 by

normal murine constituents and murine multiple myeloma cells using flow

cytometry ......................................................................................................... 108

3.4.4 N-cad protein detection by 5T33-GFP and 5TGM1-GFP cells using Western

blotting ............................................................................................................. 113

3.4.5 Protein visualisation of HSC niche molecules, ligands and CD138 by 5T33-

GFP and 5TGM1-GFP cells using immunofluorescence ................................... 115

3.4.6 Qualitative assessment of gene expression for the HSC niche molecules

and ligands by MC3T3-E1 and primary osteoblast lineage cells using endpoint

PCR ................................................................................................................... 119

Table of contents Page xiv

3.4.7 Quantitative assessment of gene expression for the HSC niche molecules

and ligands by MC3T3-E1 and primary osteoblast lineage cells using real-time

PCR ................................................................................................................... 120

3.4.8 Protein quantification of the HSC niche molecules and ligands by MC3T3-

E1 and primary osteoblast lineage cells using flow cytometry ........................ 122

3.5 Discussion .............................................................................................. 124

Chapter IV: The ex vivo and in vivo expression of HSC niche molecules and their

ligands by murine myeloma cells ..................................................................... 131

4.1 Introduction ............................................................................................... 132


4.2.1. Hypothesis and aims .............................................................................. 135

4.2.2 Objectives ................................................................................................ 135

4.3 Methods ................................................................................................ 135

4.3.1 Murine models of multiple myeloma ..................................................... 135

4.3.2 Protein analysis of bone marrow and spleen derived myeloma cells,

analysed ex vivo ............................................................................................... 136

4.3.2.1 Flow cytometry ................................................................................ 136

4.3.3 Protein analysis of bone marrow derived myeloma cells, analysed in vivo

.......................................................................................................................... 137

4.3.3.1 IHC .................................................................................................... 137

4.3.4 Statistics .................................................................................................. 138

Table of contents Page xv

4.4 Results ................................................................................................... 138

4.4.1 Quantification of tumour burden at the end-stage of disease in the bone

marrow of mice injected with 5T33-GFP and 5TGM1-GFP cells ...................... 138

4.4.2 Protein quantification of the HSC niche molecules, ligands and CD138 by

5T33-GFP and 5TGM1-GFP cells ex vivo using flow cytometry ....................... 140

4.4.3 Protein quantification of the HSC niche molecules and ligands by the bone

marrow and spleen-derived 5TGM1-GFP cells over-time analysed ex vivo using

flow cytometry. ................................................................................................ 146

4.4.4 Protein visualisation of the HSC niche molecules CXCR4 and N-cad, and

CD138 by the 5TGM1-GFP cells in histological bone sections, over-time using IHC

.......................................................................................................................... 148

4.5 Discussion .............................................................................................. 154

Chapter V: Validating a model for primary osteoblast lineage cell differentiation in

vitro, to determine the expression of HSC niche molecules and ligands, and to

develop an osteoblastic and myeloma cell co-culture system ........................... 162

5.1 Introduction ........................................................................................... 163


5.2.1 Hypothesis and aims ............................................................................... 168

5.2.2 Objectives ................................................................................................ 168

5.3 Methods ................................................................................................ 169

5.3.1 Assessment of primary osteoblast lineage cell differentiation .............. 169

5.3.1.1 Primary osteoblast lineage cell differentiation assay ...................... 169

Table of contents Page xvi

5.3.1.2 Alkaline phosphatase analysis ......................................................... 169

5.3.1.3 Mineralisation analysis ..................................................................... 170

5.3.1.4 Gene expression analysis by real-time PCR ..................................... 170

5.3.2 Assessment of 5TGM1-GFP adhesion to primary osteoblast lineage cells at

different stages of differentiation .................................................................... 171

5.3.2.1 Co-culture assessment ..................................................................... 171

5.3.3 Statistics .................................................................................................. 171

5.4 Results ................................................................................................... 171

5.4.1 Alkaline phosphatase production by primary osteoblast lineage cells

cultured in standard or osteogenic media over-time ...................................... 171

5.4.2 DNA content of primary osteoblast lineage cells cultured in standard or

osteogenic media over-time ............................................................................ 172

5.4.3 Mineralisation of primary osteoblast lineage cells cultured in osteogenic

media over-time ............................................................................................... 174

5.4.4 Gene expression of osteoblast differentiation markers and HSC niche

molecules and ligands by primary osteoblast lineage cells cultured in osteogenic

media over-time ............................................................................................... 176

5.4.5 Adhesion of 5TGM1-GFP cells to primary osteoblast lineage cells cultured

in osteogenic media ......................................................................................... 179

5.5 Discussion .............................................................................................. 181

Chapter VI: The effect of knocking down HSC niche molecules in myeloma cells

upon adhesion to osteoblast lineage cells in vitro and tumour burden in vivo... 192

Table of contents Page xvii

6.1 Introduction ........................................................................................... 193


6.2.1 Hypothesis and aims ............................................................................... 195

6.2.2 Objectives ................................................................................................ 196

6.3 Methods ................................................................................................ 196

6.3.1 Short hair-pin RNA knock-down strategy ............................................... 196

6.3.1.1 Puromycin killing curves .................................................................. 196

6.3.1.2 Polybrene optimisation .................................................................... 197

6.3.1.3 Lentiviral green fluorescent protein transduction ........................... 197

6.3.1.4 Lentiviral short hair-pin RNA transduction for CXCR4 and N-cad .... 197

6.3.2 Assessment of gene knock-down ............................................................ 198

6.3.2.1 Real-time PCR ................................................................................... 198

6.3.2.2 Western blotting .............................................................................. 198

6.3.3 Assessment of the in vitro effects of HSC niche molecule gene knock-down

in transduced myeloma cells ........................................................................... 199

6.3.3.1 Growth curves .................................................................................. 199

6.3.3.2 Adhesion experiments ..................................................................... 199

6.3.4 Assessment of the in vivo effects of HSC niche molecule gene knock-down

in transduced myeloma cells ........................................................................... 199

6.3.4.1 Tumour burden analysis................................................................... 200

6.3.4.2 Bone disease analysis ....................................................................... 200

Table of contents Page xviii

6.3.5 Statistics .................................................................................................. 201

6.4 Results ................................................................................................... 201

6.4.1 Killing curves to determine the optimum concentration of puromycin using

flow cytometry ................................................................................................. 201

6.4.2 The effect of polybrene upon cell viability after 24 hours of treatment 204

6.4.3 Green fluorescent protein transduction in the 5T33-WT cells as a proof of

principal ............................................................................................................ 205

6.4.4 Real-time PCR to establish the success of knocking-down CXCR4 and N-cad

using the lentiviral short hair-pin RNA transduction system ........................... 207

6.4.5 Real-time PCR to establish the stability of N-cad knock-down in cells

cultured in puromycin free media ................................................................... 210

6.4.6 Western blotting to establish the success of knocking-down N-cad using

the lentiviral short hair-pin RNA transduction system .................................... 211

6.4.7 Western blotting to establish the stability of N-cad knock-down in cells

cultured in puromycin free media ................................................................... 212

6.4.8 The effect of N-cad knock-down upon cell proliferation ........................ 213

6.4.9 The effect of N-cad knock-down upon adhesion to primary osteoblast

lineage cells in vitro .......................................................................................... 215

6.4.10 Flow cytometry to determine the effect of knocking-down N-cad in the

5TGM1-GFP cells upon myeloma cell growth in vivo ...................................... 218

6.4.11 IHC to determine the effect of knocking-down N-cad in the 5TGM1-GFP

cells upon myeloma cell growth in vivo ........................................................... 221

Table of contents Page xix

6.4.12 Micro-computed tomography to determine the effect of knocking-down

N-cad in the 5TGM1-GFP cells upon bone parameters in vivo ........................ 223

6.5 Discussion .............................................................................................. 225

Chapter VII: ..................................................................................................... 234

General discussion ........................................................................................... 234

7.1 General discussion.................................................................................. 235

7.2 Future studies ........................................................................................ 244

Appendices ...................................................................................................... 248

Appendix 1: Equipment, materials and reagents ........................................... 248

1.1 Tissue culture ............................................................................................. 248

1.2 Animal cell isolation ................................................................................... 249

1.3 RNA extraction ........................................................................................... 249

1.4 Reverse transcription ................................................................................. 249

1.5 Endpoint PCR .............................................................................................. 250

1.6 Gel electrophoresis .................................................................................... 250

1.7 Real-time PCR ............................................................................................. 250

1.8 Flow cytometry .......................................................................................... 251

1.9 Immunofluorescence ................................................................................. 251

1.10 IHC ............................................................................................................ 252

1.11 Western blotting ...................................................................................... 253

1.12 Primary osteoblast lineage cell differentiation ........................................ 254

Table of contents Page xx

1.13 Gene knock-down using short hair-pin RNA ............................................ 254

1.14 Micro-computed tomography ................................................................. 254

Appendix 2: Additional information for methodologies ................................ 255

2.1 RNA quantification and contamination analysis ........................................ 255

2.2 RNA fragment length analysis .................................................................... 255

2.3 Endpoint PCR primer information .............................................................. 256

2.4 Endpoint PCR primer sequence information ............................................. 256

2.5 Probes used for real-time PCR ................................................................... 257

2.6 Real-time PCR probe efficiencies ............................................................... 258

2.7 Optimisation experiments to determine the optimum amount of protein

required to visualise N-cad using Western blotting ........................................ 259

2.8 Alkaline phosphatase analysis .................................................................... 260

References ...................................................................................................... 261

Abstract Page xxi

Abstract

Multiple myeloma is an incurable B cell malignancy characterised by the expansion

of malignant plasma cells in the bone marrow. It has been suggested that during

initial colonisation of bone and possibly during therapy, some myeloma cells may

occupy a bone marrow niche similar to that inhabited by haemopoietic stem cells.

Haemopoietic stem cells residing in BM niches adhere to osteoblastic cells via a series

of molecules that promote haemopoietic stem cell quiescence. Therefore, we

hypothesise that myeloma cells express the same molecules as haemopoietic stem

cells, chemokine C-X-C-motif-receptor 4, notch-1, tyrosine kinase-2 and n-cadherin,

which interact with their complementary ligands expressed by osteoblastic cells,

chemokine C-X-C-motif-ligand 12, jagged-1, angiopoietin-1 and n-cadherin. These

interactions may result in myeloma cell adhesion to an osteoblastic niche, resulting

in myeloma cell dormancy. The aims of these studies were to determine the

expression of haemopoietic stem cell niche molecules and ligands by murine

myeloma cell lines and osteoblastic cells and to determine the role of one of the key

molecules in vitro and in vivo.

5T33MMvt and 5TGM1 cells expressed the haemopoietic stem cell niche molecules;

chemokine C-X-C-motif-receptor 4, notch-1, tyrosine kinase-2 and n-cadherin and

the MC3T3-E1 cells and primary osteoblast lineage cells expressed the ligands

chemokine C-X-C-ligand 12, jagged-1, angiopoietin-1 and n-cadherin. Knock-down of

n-cadherin was achieved in the 5TGM1 cells, with 71% gene and 75% protein

reduction. 5TGM1 n-cadherin knock-down cell attachment to primary osteoblast

lineage cells was reduced in vitro, though this did not reach significance. Mice

injected with 5TGM1 n-cadherin knock-down cells had significantly less tumour in

vivo compared to controls.

In conclusion, murine myeloma cells expressed the same repertoire of molecules as

haemopoietic stem cells and osteoblastic cells expressed their complementary

ligands. Knock-down of one of these key molecules, n-cadherin, did not significantly

Abstract Page xxii

inhibit myeloma cell attachment to primary osteoblasts in vitro but potentially

impaired tumour growth in vivo. Further experiments are required to confirm this.

Abbreviations Page xxiii

Abbreviations

ADAM A disintegrin and metalloproteinases

ADH-1 Adherin 1

AID Activation-induced deaminase

Akt Protein kinase B

Alp Alkaline phosphatase

Ang Angiopoietin

ANOVA Analysis of variance

APC Allophycocyanin

APS Ammonium persulphate

Asc Ascorbic acid

ATP Adenosine triphosphate

ATCC The American Type Culture Collection

BAD B cell lymphoma-2 associated death promoter

BCA Bicinchoninic acid

Bcl B cell lymphoma

BCSH British Committee for Standards in Haematology

BD Becton Dickinson

BFU Burst forming unit

BGP Beta-glycerophosphate

BM Bone marrow

Abbreviations Page xxiv

BMP Bone morphogenetic protein

BMPR Bone morphogenetic protein receptor

BP Base Pair

BSA Bovine serum albumin

BV/TV Bone volume/ tissue volume

CAR Chemokine C-X-C motif ligand 12 abundant reticular

CBF1α Core binding factor 1 alpha

CD Cluster of differentiation

cDNA Complementary deoxyribonucleoside

CFU Colony forming unit

CFUb Colony forming unit basophil

CFUe Colony forming unit erythrocyte

CFUEo Colony forming unit eosinophil

CFUG Colony forming unit granulocyte

CFUGEMM Colony forming unit granulocyte erythroid monocyte

and megakaryocyte

CFUGM Colony forming unit granulocyte monocyte

CFUGMEo Colony forming unit granulocyte monocyte eosinophil

CFUM Colony forming unit monocyte

CFUmega Colony forming unit megakaryocyte

Cm Centimetre

COL1A1 Collagen type I alpha I

Abbreviations Page xxv

CSL Core binding factor 1/ Recombination signal-binding

protein 1 for J-kappa/ Su9H/Lag-1

Ct Cycle threshold

CT Computed tomography

CTRL Control

Ct.th Cortical thickness

Cu Copper

CXCL Chemokine C-X-C motif ligand

CXCR Chemokine C-X-C motif receptor

DAB 3,3'-Diaminobenzidine

DAPI 4',6-Diamidino-2-phenylindole, dihydrochloride

DAPT N-[(3,5-Difluorophenyl)acetyl]-L-al-anyl-2-

phenylglycine-1,1- dimethylethyl

DEPC Diethylpyrocarbonate

Dex Dexamethasone

dH2O Distilled water

DiD 1,1′-dioctadecyl-3,3,3′,3′

tetramethylindodicarbocyanine perchlorate

DKK Dickkopf

dl Decilitre

DMEM Dulbecco’s Modified Eagle Medium

DMSO Dimethylsulphoxide

DNA Deoxyribonucleic acid

Abbreviations Page xxvi

dNTP Deoxynucleotide triphosphate

DOS Delta and oncostatin M

Dpi Dots per inch

DPX Di-N-butyl phthalate in xylene

Ds Double stranded

DSL Delta serrate lag-2 region

DTT Dithiothreitol

dUTP Deoxyuridine triphosphate

EC Endothelial cell

E-cad E-cadherin

ECL Enhanced chemoluminescence substrate

EDTA Ethylenediaminetetraacetic acid

EGF Epithelial growth factor

ERK Extracellular signal-related kinase

FACS Fluorescence-Activation Cell Sorting

FAM 6-carboxyfluorescein

FBS Foetal bovine serum

Fc Fragment crystallisable

FISH Fluorescent in situ hybridisation

FZ Fungizone

G Gram

G Gravitational acceleration

Abbreviations Page xxvii

GAPDH Glyceraldehyde 3-phosphate dehydrogenase

GDP Guanosine diphosphate

GFP Green fluorescent protein

GOI Gene of interest

GPCR Guanosine nucleotide-binding protein-coupled

receptor

GTP Guanosine triphosphate

HBSS Hank’s balanced salt solution

HCl Hydrogen chloride

HES Hairy enhancer of split

Hey1 Hairy/ enhancer of split related with YRPW motif

protein 1

HK Housekeeping

HRP Horseradish peroxidase

HPRT Hypoxanthine-guanine phosphoribosyltransferase

HRT Hairy-related

HSC Haemopoietic stem cell

HUVEC Human umbilical vein endothelial cell

H2O Water

H2O2 Hydrogen peroxide

ICAM Intercellular adhesion molecule

IF Immunofluorescence

Abbreviations Page xxviii

Ig Immunoglobulin

IGF Insulin-like growth factor

IgH Immunoglobulin heavy

IHC Immunohistochemistry

IκBα Nuclear factor of kappa light chain polypeptide gene

enhancer in B-cells inhibitor alpha

IL Interleukin

IMS Industrial methylated spirits

IP Intraperitoneal

IV Intravenous

Jag-1 Jagged-1

kb Kilo base

KCl Potassium chloride

KD Knock-down

KO Knockout

KSL C-kit+ Sca-1+ Lineage-

Kv Kilo volt

LEPR Leptin receptor

l Litre

LT-HSC Long-term haemopoietic stem cell

M Molar

MAPK Mitogen-activated protein kinase

Abbreviations Page xxix

MC Myeloma cell

MEMα Minimum essential medium alpha

MEMα-Nuc Minimum essential medium alpha containing

nucleosides

MFI Mean fluorescent intensity

mg Milligram

MGB Minor groove binder

MgCl2 Magnesium chloride

MGUS Monoclonal gammopathy of unknown significance

MHC Major histocompatibility complex

Min Minute

MIP Macrophage inflammatory protein

miRNA Micro ribonucleic acid

ml Millilitre

mm Millimetre

mM Millimolar

MM Multiple myeloma

MNNL Module at the N-terminus of Notch ligands

MOI Multiplicity of infection

MRI Magnetic resonance imaging

mRNA Messenger ribonucleic acid

ms Millisecond

Abbreviations Page xxx

MSC Mesenchymal stem cell

NaCl Sodium chloride

NaOH Sodium hydroxide

NaP Sodium pyruvate

N-cad Neural cadherin

NCKD N-cadherin knock-down

NEAA Non-essential amino acid

NF-κB Nuclear factor kappa B

ng Nanogram

NK Natural Killer

nm Nanometre

nmol Nanomolar

NOD/SCID Non-obese diabetic/severe combined

immunodeficiency

NP-40 Nonyl phenoxypolyethoxylethanol

Ocn Osteocalcin

OD Optical density

OGM Osteogenic media

OLC Osteoblast lineage cell

On Osteonectin

OPG Osteoprotegerin

Opn Osteopontin

Abbreviations Page xxxi

OSM Oncostatin-M

Ostx Osterix

PAC Puromycin N-acetyle-transferase

PBS Phosphate buffered saline

PBST Phosphate buffered saline containing tween

PCR Polymerase chain reaction

PE Phycoerythrin

PEST Proline, glutamic acid, serine and threonine

PET Positron emission tomography

PET-CT Positron emission tomography-computed tomography

PFA Paraformaldeyhyde

PI Propidium iodide

PIC Protease inhibitor cocktail

PIK-3 Phosphatidylinositol 3-kinase

PK Protein kinase

PKH Paul Karl Horan

PLC Phospholipase-c

PNPP p-nitrophenyl phosphate

POI Protein of interest

PPR Parathyroid-related protein receptor

PS Penicillin/ streptomycin

PTH Parathyroid hormone

Abbreviations Page xxxii

PTHrP Parathyroid hormone- related protein

Q Quencher

R Reporter

RAG-1 Recombination activation gene-1

RAM Recombination signal-binding protein 1 for J-kappa-

associated molecule

RANK Receptor activator of nuclear factor kappa beta

RANKL Receptor activator of nuclear factor kappa beta ligand

RBC Red blood cell

RBP-JK Recombination signal-binding protein 1 for J-kappa

RISC Ribonucleic acid induced silencing complex

RNA Ribonucleic acid

RNAi RNA interference

RNasin Ribonucleic acidase inhibitor

RPM Rotations per minute

RPMI Roswell Park Memorial Institute medium

RT Reverse transcription

RT- Complementary deoxyribonucleic acid samples

without reverse transcriptase enzyme

RT+ Complementary deoxyribonucleic acid samples with

reverse transcriptase enzyme

RVD Lenalidomide, bortezomib, dexamethasone

SA Streptavidin

Abbreviations Page xxxiii

SA-HRP Streptavidin horseradish peroxidase

SC Subcutaneous

SCF Stem cell factor

SCID Severe combined immunodeficiency

SDF Stromal-derived factor

SDS Sodium dodecyl sulphate

SDS-PAGE Sodium dodecyl sulphate-polyacrylamide gel

electrophoresis

SEM Standard error of the mean

SFRP Secreted frizzled-related protein

shRNA Short hair-pin RNA

siRNA Silencing ribonucleic acid

SNP Single nucleotide polymorphism

ST-HSCs Short-term haemopoietic stem cell

TAD Transcriptional activation domain

TAE Trizma base, acetic acid and

ethylenediaminetetraacetic acid

Tb.n Trabecular number

Tb.pf Trabecular pattern factor

Tb.th Trabecular thickness

TEMED Tetramethylethylenediamine

TGF Transforming growth factor

Abbreviations Page xxxiv

Tie Tyrosine kinase

TIFF Tagged image file format file

Tm Melting temperature

TNF Tumour necrosis factor

TO-PRO-3 TO-PRO-3 iodide

TRAIL Tumour necrosis factor-related apoptosis-inducing

ligand

TU Transducing units

UD Undetermined

UV Ultra violet

VCAM Vascular cell adhesion molecule

V (D) J Variable, diversity, joining

VEGF Vascular endothelial growth factor

VLA-4 Very late antigen

WB Western blotting

WBC White blood cell

Wnt Wingless

WT Wild type

Zn Zinc

5-FU Fluorouracil

5’ Five prime end

5TGM1-GFP 5TGM1-green fluorescent protein

Abbreviations Page xxxv

5TGM1vt 5TGM1 in vitro

5T33-GFP 5T33MM-green fluorescent protein

5T33MMvt 5T33 multiple myeloma in vitro

5T33-GFP 5T33MM-green fluorescent protein

3’ Three prime end

Α Alpha

Β Beta

β2M Beta 2 microglobulin

°C Degrees Celsius

Δ Delta

ΔCt Delta cycle threshold

γ Gamma

κ Kappa

∏ Pi

µA Microamperes

µCT Micro computed tomography

µg Microgram

µl Microlitre

µm Micrometre

µmol Micromolar

Chapter I: Introduction Page 1

Chapter I: Introduction


1.1 Multiple Myeloma

1.1.1 Epidemiology of multiple myeloma

Multiple myeloma (MM) is a haematological malignancy predominantly

characterised by the uncontrolled expansion of plasma cells in the bone marrow

(BM), resulting in bone disease (1). It accounts for 1.5% of all cancers (2) and for at

least 13% of all haematological malignancies worldwide (3). In the UK, the incidence

of MM is 6-8 diagnoses per 100,000 people with over 4500 people diagnosed with

MM per year (2).

MM is an age-dependent disorder where the mean age of diagnosis is 62 years for

men (75% >70 years) and 61 years for women (79% >70 years) (4) and less than 15%

are diagnosed before the age of 60 (5). It is the most common haematological cancer

in African Caribbean ethnicities where twice as many are likely to develop myeloma

compared to European Americans (6). There is also a higher incidence of MM in

males rather than females (7). However, the genetic basis for any differences

between specific ethnicities and sexes is unknown. When the frequency of

immunoglobulin (Ig) heavy (IgH) translocations, translocation loci and genomic gain

or loss were compared between European Americans and African Caribbeans, the

only statistical difference observed was a greater frequency of IgH translocations in

European Americans (6).

1.1.2 Pathology of disease

Myeloma cells (MCs) are classified as abnormal plasma cells. The primary role of a

normal plasma cell is to provide a humoral immune response via the production of

Igs when a foreign antigen is encountered. Plasma cells are typically un-dividing cells,

which possess a distinct morphology, high endoplasmic reticulum and Golgi content,

as well as the expression of specific molecules. Normal plasma cells express

molecules such as syndecan-1 (also known as cluster of differentiation (CD) 138),

very-late antigen (VLA)-4 and chemokine C-X-C motif receptor (CXCR) 4. Markers that


are down-regulated post plasma cell differentiation include B220, CD19, CD21, CD22,

CXCR5 and the major histocompatibility complex (MHC) II (8).

Plasma cells originate from B lymphocytes which are produced from differentiated

BM haemopoietic stem cells (HSCs), discussed in Chapter I Section 1.2. Throughout

the stages of differentiation, several significant genetic modifications occur, which

include; rearrangements of the variable, diversity and joining (V(D)J) regions of the

IgH chain during somatic hypermutation (8, 9) and Ig class switching in the centrocyte

and plasmablast (10). Both somatic hypermutation and class switching use

activation-induced deaminase (AID), which produces double strand deoxyribonucleic

acid (DNA) breaks. The DNA breaks are usually repaired locally however, they may

lead to chromosomal translocations, which can occur in up to 1000 cells a day (11).

Therefore, each DNA modification that occurs within a cell throughout the stages of

B cell differentiation increases the risk of chromosomal abnormalities, which are a

common feature in myeloma (12).

Several genetic mutations have been identified in myeloma patients and they are

characterised into two cytogenetic categories, which include hyperdiploid or non-

hyperdiploid. Hyperdiploid patients have more than the normal diploid number of

chromosomes such as trisomies of chromosomes (this may include chromosomes 3,

5, 7, 9, 11, 15, 19), whereas non-hyperdiploid patients are hypodiploid, with fewer

than the normal diploid number of chromosomes or may have chromosomal

translocations, particularly of the IgH gene, which occurs in approximately 50% of

monoclonal gammopathy of unknown significance (MGUS) patients (the pre-

malignant condition before asymptomatic or symptomatic myeloma) and 60-65% of

symptomatic myeloma patients (13, 14).

1.1.3 Clinical features

MM patients can present with several different clinical features, depending on the

stage and severity of disease. Common features are haematological irregularities,

renal disease and bone disease.

Haematological irregularities include a reduction in erythrocytes, platelets and white

blood cells (WBCs) as well as normochromic and normocytic anaemia (iron <8.5

grams (g) per decilitre (dl)) which are exhibited as fatigue, abnormal bleeding and

increased risk of infection. These signs are primarily caused by the rapid expansion

of MCs in the BM (15).

Renal disease is also common in MM patients and is “one of the most common causes

of death, second to infection” (16). Up to 50% of newly diagnosed patients can

present with renal abnormalities and 20% of patients are diagnosed with severe renal

impairments which require dialysis (17). Renal failure is caused by the presence of

malignant plasma cells in the BM. Clonal plasma cells secrete monoclonal antibody,

also known as paraprotein or M protein (50% of patient MCs secrete IgG, 20% IgA

and 20% light chains) (17), which cause lesion formation in the kidneys, known as

cast nephrology. Lesions present in the kidneys have been associated with tubular

rupture and nephritis, resulting in obstructions and poor filtration (16). In addition,

amyloidosis caused by paraprotein or light chain secretion as well as hypercalcaemia,

caused by the breakdown of bone (discussed in more detail in Chapter I section 1.1.5)

also result in renal impairment (2).

Bone disease and the development of osteolytic lesions, as a result of MC infiltration

of the BM, are also a major cause of morbidity and occur in 40-50% of MM patients

(18). Patients suffering from lesions may experience bone pain and have increased

risk of pathological fractures (2), discussed in greater detail in Chapter I, Section

1.1.5.


1.1.4 Diagnosis

Several tests are performed to reach a diagnosis of MM. These include analysis of the

blood, urine, BM and skeleton.

Full blood counts, red blood cell (RBC) sedimentation rate and blood smears are used

to assess abnormalities in the number of different cellular constituents in the

peripheral blood, as well as to test iron and haemoglobin levels. Bone marrow

aspirates or trephines may also be taken to determine the number of different

cellular constituents, particularly the number of plasma cells, which is generally

higher than 10% in MM patients (5). Plasma cell phenotyping can also be conducted

using flow cytometry to characterise specific cell surface markers, as well as

fluorescent in situ hybridisation (FISH) to determine cytogenetic abnormalities (19).

Karyotyping can also be conducted to identify specific chromosomal abnormalities.

Renal defects are detected by urine biochemistry or urine electrophoresis, which

detect the presence of paraprotein or Ig light chains, also known as Bence Jones

protein, in the urine. In addition, serum electrophoresis, immuno-fixation and

biochemical tests are also conducted to determine the presence of paraprotein

present in the circulation as well as levels of creatinine, calcium, beta (β) 2

microglobulin (β2M) and albumin, which indicate severity of renal dysfunction (5,

20).

Positron emission tomography (PET)-computed tomography (CT) (PET-CT), magnetic

resonance imaging (MRI) and X-rays are used to visualise the extent of bone lesions

and fractures in MM patients (20).

In early MM, it is often difficult to distinguish symptomatic MM from related

disorders such as asymptomatic/smouldering MM and MGUS. Monoclonal

gammopathy of unknown significance is the most common premalignant condition

present in Western countries and is present in 3% of the general population over 50

years of age (21, 22). It is diagnosed by the presence of paraprotein however, levels

are less than 30 g/l and patients do not show signs of bone disease or organ-related

disorders. The risk of developing MM or plasma cell related disorders increases at a

rate of 1% every year following diagnosis, and it is uncertain whether all patients

have the premalignant condition prior to MM diagnosis. However, Landgren et al (22)

found that “virtually all” MM patients in their studies (N=71) had a form of MGUS

(diagnosed by the presence of paraprotein < 30 g/dl) prior to MM, but the trigger

which initiates progression to MM is unknown.

Smouldering/ asymptomatic MM is a precursor to MM, whereby approximately 3000

cases are diagnosed per year in the United States of America and patients have an

average annual risk of progression to MM of 10% in the first 5 years after diagnosis,

which decreases to 3% annually for the following 5 years and 1% thereafter (23).

Smouldering MM patients are diagnosed due to paraprotein levels above 30 g/dl and

more than 10% of plasma cells present in the BM, however they do no present with

any organ related impairment (5).

An international diagnostic system to distinguish between MGUS, smouldering MM

and symptomatic MM was developed by Smith et al (24), the International Myeloma

Working Group (25) and the British Committee for Standards in Haematology (BCSH),

as stated in the official Myeloma Guidelines (5), shown in Table 1.1.

Table 1.1: The diagnostic criteria to differentiate between MGUS, asymptomatic MM and symptomatic MM

Adapted from the Bird et al (5), permission from The British Journal of Hematology 04.08.14. l: Litre.

In addition, a staging system was also designed by Durie and Salmon (26) and has

been adapted by the BCSH in the MM guidelines (5), as shown in Table 1.2. This

Monoclonal gammopathy of

unknown significance Asymptomatic myeloma Symptomatic myeloma

M protein in serum <30 g/l

M protein in serum >30 g/l

M protein present in serum and/or urine

BM clonal plasma cells <10%

And/or BM clonal plasma cells >10%

BM plasma cells and plasmacytoma

No myeloma related organ or tissue

impairment

No myeloma related organ or tissue

impairment, bone lesions or symptoms

Any myeloma related organ or tissue

impairment including bone lesions

staging system was designed to categorise patient stage and severity of disease in

order to reach a prognosis and to decide upon suitable treatment options.

Table 1.2: The Durie and Salmon staging system of myeloma

Adapted from Durie and Salmon (26) by the BCSH (5), permission from The British Journal of Hematology 04.08.14. nmol: nanomole, µmol: micromole, mg: milligram.

1.1.5 Pathology of tumour-induced lytic bone disease

The specific constituents of the BM and its microenvironment provide a suitable

“soil” for the expansion and enhanced survival of MCs in MM, as explained by Paget

(27) using the “seed and soil” hypothesis. Two specialised cells residing in the BM

microenvironment, which are of importance with regards to the development of

MM, are osteoblasts and osteoclasts. Osteoblasts are responsible for secreting the

extracellular organic bone matrix known as osteoid, which is then mineralised to

form new bone (28). Osteoclasts are able to secrete enzymes and proteins to digest

the bone in the process of resorption (29). The mechanisms in which osteoblasts and

osteoclasts work together to produce functioning bone is known as remodelling (30)

and this balanced process is described in Figure 1.1.

Stage Criteria Median survival (months)

Stage I Serum β2M < 3.5 mg/l (296 nmol/l)

Serum albumin ≥ 3.5 g/dl

(35g/l or 532 µmol/l)

62

Stage II Remainder fitting neither Stage I or III

45

Stage III Serum β2M ≥ 5.5 mg/l (465 nmol/l)

29


Figure 1.1: A schematic diagram representing the cycle of normal bone remodelling. The cycle consists of four phases: 1. Bone resorption occurs by the activation of osteoclasts, which release acid phosphatases and collagenases, which degrade bone. 2. The reversal phase in which apoptosis of osteoclasts occurs followed by recruitment of osteoblasts. 3. Subsequent bone formation where osteoblasts secrete osteoid, which is mineralised with hydroxyapatite to form new bone. 4. A resting phase in which neither resorption or formation occurs and osteoblastic cells become flattened bone lining cells (31).

In MM, the process of bone remodelling is “uncoupled” (32) and the

osteoclast/osteoblast axis is disrupted, favouring osteoclastic bone resorption and

supressing osteoblastic bone formation, ultimately leading to lytic bone lesions and

bone disease (18). This increase in bone loss is enhanced by what is known as “The

Vicious Cycle” (33), where myeloma cells express or secrete factors such as nuclear

factor kappa (κ) β ligand (RANKL) (32, 34), macrophage inflammatory protein 1

(MIP1α) (35, 36), tumour necrosis factor α (TNFα) (37), interleukin 1 β (IL-1β) (38), IL-

3 (39) and IL-6 (40), which stimulate osteoclastic differentiation and activation. In-

turn, osteoclasts produce proteins including IL-6 (41), annexin II (42), osteopontin

(43), B cell-activating factor belonging to the TNF family (BAFF) (44) and a

proliferation-inducing ligand (APRIL) (44), which increase myeloma cell proliferation

or prevent apoptosis. Overall, this cycle results in increased numbers of active

osteoclasts and myeloma cells, leading to enhanced bone loss. This phenomenon is

illustrated in Figure 1.2.


Figure 1.2: A schematic diagram representing “The Vicious Cycle”. Myeloma cells produce the factors: RANKL, MIP1α, TNFα, IL-3 and IL-6 that stimulate osteoclast differentiation. In-turn, osteoclasts produce: IL-6, annexin II, osteopontin, BAFF and APRIL, which induce myeloma cell proliferation. These events occur within a cycle, driving both myeloma cell expansion and osteoclastogenesis resulting in a net loss of bone. Adapted from Galson et al (33), permission from Nature Reviews 31.07.14.

One of the main regulators of osteoclastogenesis in MM is the RANK/RANKL

pathway. Receptor activator of nuclear factor κ β ligand is an essential cytokine for

the differentiation of preosteoclasts to activated osteoclasts (45). Receptor activator

of nuclear factor κ β ligand is expressed by stromal cells, osteoblasts and T cells and

exists as a membrane bound form as well as soluble RANKL, when cleaved by

metalloproteinases. Receptor activator of nuclear factor κ β ligand binds to its

receptor RANK, expressed on preosteoclasts and activated osteoclasts. Receptor

activator of nuclear factor κ β / RANKL interactions result in the differentiation of

preosteoclasts, the fusion of polykaryons and the activation of resorption in mature

osteoclasts. Osteoclastogenesis can however, be inhibited by a decoy receptor,

known as osteoprotegerin (OPG), which is expressed by osteoblasts. Osteoprotegerin

binds to RANKL, blocking the binding of RANK with RANKL, thus behaving as a

paracrine inhibitor of osteoclast formation and activation (46). In MM, OPG is down

regulated (47) therefore, osteoclastogenesis and thus bone resorption is not

inhibited resulting in a disruption of normal bone remodelling.


As mentioned above, osteoblast number and bone formation are reduced in MM

which may be due to inhibition of the wingless (Wnt) signalling pathway by MC

secretion of soluble Wnt inhibitory factors including dickkopfs (DKKs) (48, 49) or

secreted frizzled-related proteins (SFRPs) (50).

1.1.6 Treatment options

To date, MM remains an incurable disease and resulted in approximately 10,710

deaths in the USA in 2012 (51) and approximately 2600 in the UK in 2011

(http://www.cancerresearchuk.org/cancer-

info/cancerstats/keyfacts/myeloma/myeloma). Survival rates post-diagnosis vary

from several weeks to several decades. However, the 5 year or more survival rate in

myeloma patients in the USA between 2002-2004 was 45-57% in patients less than

50 years of age, 39-48% in 51-59 year olds, 31-36% in 60-69 year olds and only 27-

29% in those over the age of 70 (52). This was similar to UK statistics in 2010 where

only 37% of patients survived 5 years post diagnosis

(http://www.cancerresearchuk.org/cancerinfo/cancerstats/types/myeloma/surviva

l/multiple-myeloma-survival-statistics#1_5_10_yr_survival).

There are a variety of treatments available for MM patients however, current

treatments are not curative, but alleviate the symptoms to improve quality of life.

Current treatments include radiotherapy, chemotherapy, bisphosphonates, surgery

and BM and stem cell transplants. The type of treatment administered is dependent

on the age of the patient and the severity of disease.

1.1.7 Treatments to reduce tumour burden

Radiotherapy is generally administered locally at the site of pain to reduce tumour

burden which in-turn reduces osteoclastic bone resorption to help prevent further

bone loss. Radiotherapy is also used to treat spinal cord compression and this may

be given in combination with surgery or kyphoplasty (53).

http://www.cancerresearchuk.org/cancer-info/cancerstats/keyfacts/myeloma/myeloma

http://www.cancerresearchuk.org/cancer-info/cancerstats/keyfacts/myeloma/myeloma

http://www.cancerresearchuk.org/cancerinfo/cancerstats/types/myeloma/survival/multiple-myeloma-survival-statistics#1_5_10_yr_survival

http://www.cancerresearchuk.org/cancerinfo/cancerstats/types/myeloma/survival/multiple-myeloma-survival-statistics#1_5_10_yr_survival


Chemotherapeutic drugs are administered to reduce tumour burden. These include

chemotherapies (lenalidomide, pomalidomide, thalidomide), corticosteroids

(dexamethasone (dex) and prednisone), and proteasome inhibitors (bortezomib and

carfilzomib). Treatments are given in combination such as lenalidomide, dex and

bortezomib (RVD), lenalidomide cyclophosphamide and dex or RVD plus

cyclophosphamide (54). Lenalidomide, thalidomide and pomolidomide are thought

to work by inducing apoptosis in MCs via caspase 8 and tumour necrosis apoptosis-

inducing ligand (TRAIL) signalling as well as down regulation of nuclear factor κ B (Nf-

κB), IL-6 and vascular endothelial growth factor (VEGF) (54-56). Proteasome

inhibitors such as bortezomib and carfilzomib are thought to act by inhibiting the 26

S proteasome, resulting in the inhibition of the degradation of proteins involved in

cell proliferation (57). In addition, bortezomib was found to inhibit angiogenesis by

reducing the production of VEGF, IL-6, IGF-1, Angiopoietin (Ang)-1 and Ang-2 by

endothelial cells in vitro (58), which could contribute to a reduction in tumour burden

in patients.

1.1.8 Treatments to alleviate symptoms

Bisphosphonates are the most commonly administered drug to reduce bone

resorption and to prevent further formation of osteolytic lesions or osteopenia.

There are several different bisphosphonates available which include the nitrogen-

containing bisphosphonates; zoledronic acid, pamidronate and ibandronate and

non-nitrogen containing bisphosphonates; clodronate and etidronate.

Bisphosphonates once administered, bind to hydroxyapatite in the bone and are

released during bone resorption and internalised by osteoclasts. Once inside the cell

they can either inhibit osteoclastic activity or induce osteoclast apoptosis (59). Non-

nitrogen containing bisphosphonates are able to inhibit adenosine triphosphate

(ATP)-dependent enzymes (60), whereas nitrogen containing bisphosphonates

inhibit the enzyme farnesyl pyrophosphate synthase, required in the mevalonate

pathway, an essential pathway for the post-translational modification of proteins

(61). In the clinic, standard therapy usually uses zoledronic acid or pamidronate to


reduce bone pain (62). In addition, bisphosphonates have also been reported to exert

an anti-tumour effect in vitro (63) and in vivo (64).

Autologous stem cell transplants may also act as standard therapy for newly

diagnosed patients, however, transplants are usually only applied to younger

patients and use also depends on the stage of disease. Other therapies that may be

required to alleviate symptoms include dialysis, rehydration, blood transfusion,

plasmaphoresis, antibiotics and Ig infusions (15).

1.1.9 Recurrence of disease

Despite continual development of novel therapies for MM, current treatments are

unable to completely eradicate the disease and patient relapse is inevitable. Previous

studies have proposed that residual MCs present in the BM following treatment are

able to evade chemotherapy via (currently) unknown mechanisms; and some groups

have categorised these cells as “cancer stem cells”(65). Several mechanisms have

however been proposed. For example, previous studies have suggested that chemo-

evasion is a result of MC mediated alterations in immune surveillance (suppression

and dysfunction of T cells and natural killer (NK) cells) (66, 67); a lack of optimal

microenvironment for growth, and some have also suggested that MC adhesion to

specific niches in the BM microenvironment results in MC dormancy (65, 68).

Concerning the latter, previous reports have speculated that dormant MCs may in

fact hijack or reside in similar niches to those occupied by HSCs and they therefore

utilise similar mechanisms as HSCs to acquire cell dormancy. The concept of the HSC

niche has been explored for approximately the last thirty-five years and previous

studies have found that HSCs were maintained in these niches by the expression and

interactions of specific molecules on the HSCs and other cells in the BM

microenvironment (69) (discussed in more detail below). In support of a MC niche, a

recent study by Chen et al (70) found early MCs (identified by a lipophilic dye known

as Paul Karl Horan dye (PKH26)) preferentially resided in osteoblastic niches and

homed and survived to a lesser extent in vascular and splenic niches, using murine

MM in vivo models. In addition, PKH26 positive cells demonstrated increased


chemoresistance compared to PKH26 negative cells. in vitro (70). However, current

data is still inconclusive as to the specific molecule interactions responsible for the

adhesion of MCs to these niches and mechanisms of MC dormancy and therefore,

further research is required to explore this.

1.2 The HSC Niche

1.2.1 Evidence for a HSC niche

HSCs possess the ability for unlimited self-renewal and production of a variety of

differentiated progeny by haematopoiesis (71), as illustrated in Figure 1.3. Schofield

et al (72) proposed that HSCs reside in a specialised microenvironment or niche which

is responsible for controlling self-renewal and differentiation of HSCs, required for

maintenance of the different cell types within the BM. The HSC microenvironment is

comprised of several supporting cells and components including stromal cells,

extracellular matrix molecules and growth factors (73) all of which contribute to the

maintenance of the stem cell pool size.


Figure 1.3: HSC differentiation into different cell lineages. A schematic diagram representing the process of haematopoiesis and the lineages in which each cell in the BM is derived. Haemopoietic stem cells differentiate into lymphoid and myeloid lineages. The lymphoid lineage produces B cells, T cells and NK cells. The myeloid lineage produces several progenitors which include colony forming unit (CFU), granulocyte erythroid monocyte and megakaryocyte (CFUGEMM), CFU granulocyte monocyte eosinophil (CFUGMEo), CFU eosinophil (CFUEo), CFU granulocyte monocyte (CFUGM), CFU granulocyte (CFUG), CFU monocyte (CFUM), CFU basophil (CFUB), CFU megakaryocyte (CFUMega), burst forming unit (BFU) and CFU erythrocyte (CFUE). These progenitors then differentiate into their corresponding mature cell type. Adapted from Hoffbrand A.V (15) and Reya et al (74).


It has been proposed that HSCs may reside in two different niches entitled the

“osteoblastic niche” and the “vascular niche” (75, 76) which are responsible for the

maintenance of HSCs and their ability to survive and proliferate in the BM, as shown

in Figure 1.4 . However, the existence of either niche remains controversial (77).

In addition to osteoblastic cells and ECs, several other cell types have been implicated

in the HSC niche in the BM. These included the chemokine C-X-C motif ligand (CXCL)

12 abundant reticular (CAR) cells, mesenchymal stem cells (MSCs), osteoclasts and

adipocytes (76). However, evidence is limited with regards to their roles in the HSC

niche.

Figure 1.4: The HSC niche. A schematic diagram illustrating the HSC niche. a. Haemopoietic stem cells adhere to osteoblasts or bone lining cells on the bone surface, which has been proposed to result in HSC quiescence. b. Once activated, the HSCs are thought to mobilise to the blood vessels where they proliferate and differentiate.

1.2.2 The osteoblastic HSC niche

The osteoblastic niche is composed of osteoblasts or “osteoblastic-like” cells

possessing an osteoblastic phenotype, which includes expression of alkaline

phosphatase (Alp). It has been hypothesised that HSCs adhere and interact with

osteoblasts or their growth factors on trabecular and endocortical surfaces of bone

(73), resulting in G0 cycling quiescent HSCs and inhibition of their maturation.

a b


Quiescent HSCs are often termed long term-HSCs (LT-HSCs) and possess a commonly

recognised phenotype of c-kit positive, sca-1 positive and lineage negative (KSL) and

possess the ability to self-renew unlike short-term HSCs (ST-HSCs) (69).

Several in vivo studies have identified the importance of osteoblastic cells for HSC

survival and self-renewal in the BM, investigated by manipulating osteoblastic

number using transgenic mice or administration of hormones in murine models. In

core binding factor 1 α (CBF1α) (an essential osteoblast differentiation transcription

factor) knockout (KO) mice (a lethal mutation in homozygous mice), this resulted in

a lack of BM cavity, little or no ossification of the skeleton and a decrease in

osteoblast numbers. In addition, osteoblastic cells that were present seemed to be

poorly differentiated with no osteocalcin (Ocn) expression and weak Alp expression

(78). In CBF1α KO embryos, loss of CBF1α and therefore osteoblasts, correlated with

a reduction in BM haemopoietic cells and also splenomegaly as a result of

extramedullary haematopoiesis (79). These findings were also supported by Visnjic

et al (71, 80) who noted a lack of BM cellularity and a reduction in KSL HSCs when

osteoblast numbers were decreased using a thymidine kinase and 2.3 kilo base (kb)

fragment of collagen type I α I (COL1A1) promotor, conditional KO mouse model.

Osteoblast stimulation using a constitutively active parathyroid hormone (PTH)/

parathyroid hormone-related protein (PTHrP) receptor (PPR) mouse model (81)

resulted in an increase in trabecular osteoblastic cells which correlated with a

significant increase in HSCs in the long bones and increased HSC engraftment (82). In

support of this, daily administration of PTH in C57Bl/6 mice, which increased

osteoblast number, also resulted in an increase in HSC number in the BM and HSC

engraftment in vivo (82). In addition, Zhang et al (83), also noted that when

osteoblast number was increased using an inducible bone morphogenetic protein

(BMP) receptor (BMPR) 1 α KO mouse model, LT-HSC number and HSC engraftment

was also increased.

The location of HSCs in the BM microenvironment is also important when defining

the HSC niche. Xie et al (84) and Zhang et al (83) both noted HSC localisation and/or


adhesion to osteoblasts on the endosteal and trabecular bone surfaces by

immunohistochemistry (IHC) or immunofluorescence (IF). The osteoblasts that the

HSCs adhered to seemed to be a distinct sub-population of mononuclear “spindle-

shaped” osteoblasts, positive for neural cadherin (N-cad), a potentially important

molecule for HSC proximity to osteoblasts and maintenance of HSC quiescence (83).

Several studies have also reported that LT-HSCs specifically adhering to bone are less

sensitive to stresses such as treatment with the chemotherapeutic fluorouracil (5-

FU) in vivo, which is potentially facilitated by their G0 phase of cell cycle (85).

1.2.3 The vascular HSC niche

HSCs have also been reported to reside at peri-vascular surfaces in the BM,

implicating a vascular niche may also be in existence. The vascular niche is

composed of single layered ECs forming sinusoids, which lack connective tissue and

therefore are highly permeable (86). It is hypothesised that active HSCs adhere to

the ECs, establishing a site for proliferation, differentiation and mobilisation from

the niche into the blood.

To illustrate the importance of ECs with regards to HSC maintenance, several in vitro

studies have co-cultured tyrosine kinase (Tie)-2 positive murine ECs, (isolated from a

variety of murine organs) with phenotypic HSCs (87). Co-culturing ECs with HSCs

promoted HSC differentiation and cell expansion, compared to controls, illustrating

the potential for ECs to provide an environment in which HSCs can survive and self-

renew. Several in vivo mouse models have also demonstrated the importance of

sinusoids for maintenance of haematopoiesis in the liver (88) and HSCs were found

adjacent to stem cell factor (SCF) expressing vessels in the BM (89). Furthermore, KO

of SCF in both Tie-2 positive ECs and leptin receptor (LEPR) positive peri-vascular

stromal cells, significantly reduced HSC frequency in the BM and reduced BM

cellularity in the SCF-LEPR KO animals (89), demonstrating the importance of the

vasculature and peri-vasculature for the maintenance of HSC numbers.


1.3 Key molecules in the osteoblastic niche

Specific receptors expressed by HSCs and their complementary ligands expressed by

either osteoblasts or ECs, may facilitate the adhesion of HSCs to the bone or

vasculature. Adhesion to these two sites may also induce HSC dormancy or activation

as well as influencing HSC homing to the BM or mobilisation into the circulation.

Several molecules of potential importance have been identified which include CXCR4

and CXCL12, Notch-1 and Jagged (Jag)-1, Tie-2 and Ang-1 and N-cad. Throughout this

thesis CXCR4, Notch-1, Tie-2 and N-cad will be referred to as the HSC niche molecules,

whereas CXCL12, Jag-1 and Ang-1 will be referred to as the complementary ligands

to the HSC niche molecules.

1.3.1 CXCR4/CXCL12

Chemokines are a large family of cytokines associated with many chemotactic

responses, particularly inflammation. The family consists of over 50 structurally-

related proteins which are 60-88 amino acids (AA) long and they are subdivided into

four groups (90). They are classified upon their NH2 terminal cysteine structure,

which contain either adjacent cysteine AAs (CC), two cysteines separated by one

different AA (CXC), two cysteines separated by three different AAs (CX3C) (91) or

contain only one cysteine (XC) (91). The chemokine CXCL12 or stromal-cell derived

factor (SDF)-1, has two splice variants, α and β. It is expressed by several cell types

such as ECs, peri-vascular cells, (92) as well as stromal cells residing on endosteal

bone surfaces (93). CXCL12 is also expressed on CAR cells residing near sinusoids in

the BM (94).

CXCL12 binds to two receptors, CXCR4 and CXCR7 (95), which are known as hepta-

helical receptors as their chain of AAs “traverses” the membrane seven times (10)

(Figure 1.5). They are also more commonly known as G-protein coupled receptors

(GPCRs), which are the largest family of membrane proteins which “mediate cellular


responses to hormones, neurotransmitters and are also responsible for vision,

olfaction and taste” (96). CXCR4, is one of nineteen known human chemokine

receptors (97) and it is expressed in a variety of different tissues including neural,

muscle, cord blood and stem cell populations (98), specifically HSCs (99).

As a GPCR, CXCR4 activation is mediated by an “intracellular heterotrimeric G

protein” which is bound to the cell membrane (100). The G protein is comprised of

three subunits, Gα, Gβ and G gamma (γ), which bind to guanosine diphosphate (GDP)

when the ligand CXCL12 is not bound. However, when CXCL12 is bound, GDP is

replaced by guanosine triphosphate (GTP), resulting in the dissociation of the G

protein into two separate subunits containing Gα which is bound to the GTP and a

second subunit containing Gβ and Gγ (100). The Gα subunit can also be classified into

4 separate families Gαs, Gαi, Gαq, and Gα12 (101, 102). Each Gα, and Gβ and Gγ

subunit, transmits the GPCR signal down different signalling pathways, which include

phospholipase-c (PLC), phosphatidylinositol 3-kinase (PI3-K), extracellular signal-

related kinase (ERK), protein kinase (PK) B (PKB/Akt), NF-κB, mitogen-activated

protein kinases (MAPK) and B cell lymphoma (bcl)-2 associated death promoter

(BAD) pathways. This results in different downstream effects, which may include

chemotaxis, gene transcription and cell survival, required for wound healing,

angiogenesis, metastasis, cell recruitment and inflammation, to name but a few (94,

100, 103-107) (Figure 1.5). CXCR4 and CXCL12 signalling is also crucial for functioning

tissues, as demonstrated by Nagasawa et al, who showed that KO of CXCL12 in a

transgenic model resulted in foetal mortality and specifically defects in haemopoeitic

cells, cardiac tissue and brain tissue (108) .


Figure 1.5: The structure of CXCR4 and activation with CXCL12. a. Diagram illustrating the general structure of CXCR4 in its unstimulated state, consisting of a hepta-helical structure bound to a G protein consisting of α, β, γ subunits bound to GDP. b. Activation of CXCR4 by CXCL12, causes the replacement of GDP with GTP, resulting in the dissociation of the Gα with Gβ and Gγ. c. Gα consists of 4 subfamilies, Gαi, p, s and 12. The α subfamilies and Gβ and Gγ subunit stimulate different signalling pathways resulting in different functions. Adapted from Sodhi et al (102), permission from Nature Reviews 30.7.14.

The requirement for CXCR4/CXCL12 in HSC homing and mobilisation have been

illustrated using CXCL12 (109) and CXCR4 KO (94) mouse models. Ara et al (109)

showed that KO of CXCL12 in mouse embryos resulted in a reduction of HSCs in the

liver, spleen and BM. In addition, Tzeng et al (110) demonstrated that conditional KO

of CXCL12 resulted in a reduction in LT-HSCs, an increase in ST-HSCs and fewer G0

cycling HSCs, therefore, demonstrating that the CXCR4/CXCL12 interactions may

promote HSC quiescence.

These results were also consistent with the conditional KO of CXCR4, where fewer

LT-HSCs and fewer G0 cycling HSCs were present in the BM and there was also an


increase in HSCs in the peripheral blood (94). Therefore, in addition the promoting

quiescence, the CXCR4/CXCL12 axis is required for HSC homing to the BM, liver or

spleen from the circulation.

CXCR4/CXCL12 signalling can be inhibited using antibodies or the antagonist

AMD3100. In vitro migration assays demonstrated that HSCs migrate in response to

CXCL12-α enriched media (99) and this was inhibited upon treatment with an anti-

CXCR4 antibody (111). Pre-treatment of CD34 positive HSCs with anti-CXCR4 or anti-

CXCL12 antibodies also severely reduced CD34 positive HSC engraftment in the BM

when injected into non-obese diabetic/severe combined immunodeficiency

(NOD/SCID) mice (112). This therefore again, reiterates the importance of

CXCR4/CXCL12 signalling for the migration and homing of HSCs.

1.3.2 Notch-1/Jag-1

Notch-1 and Jag-1 have both been reported to possess several roles in stem cell fate,

self-renewal and differentiation into specific progeny from HSCs (113, 114). Notch

molecules exist as a family of transmembrane glycoprotein receptors, which occur as

four homologues, 1-4 (115). Their basic structure consists of an intracellular and an

extracellular domain, which contain several different motifs (116-118) as shown in

Figure 1.6.

Notch-1 is expressed in a variety of different organs and tissues, such as the thymus,

BM and foetal liver (119) as well as by various cell types which include BM precursors,

lymphocytes, neutrophils (115) and CD34 positive HSCs (120).


Figure 1.6: The structure of Notch-1. A diagram demonstrating the extracellular and intracellular structure of Notch-1. Its extracellular region consists of epidermal growth factor (EGF) like proteins and Notch-1/ Lin-1 repeats and its intracellular domain consist of a recombination signal-binding protein 1 for J-κ (RBP-JK)-associated molecule (RAM), ankyrin repeats and a transcriptional activation domain (TAD) and a proline, glutamic acid, serine and threonine (PEST) region. Adapted from Swiatek et al (121).

Notch-1 is activated by its complementary ligand Jag-1. The ligands were first

characterised in Drosophila, which consist of two main groups of Notch ligands

named Serrate and Delta. The mammalian homologues of each of these groups of

ligands are Delta 1 and 2 and Jag-1 and 2 (113). Jag-1, illustrated in Figure 1.7, is

expressed on several cell types, including primary BM stromal cells and human

umbilical vein ECs (HUVECs) (114) as well as the murine preosteoblastic cell line

MC3T3-E1 (122).


Figure 1.7: The structure of Jag-1. A diagram illustrating the structure of Jag-1, which consists of an extracellular domain containing a cysteine rich region, EGF-like repeats, a delta (Δ) and oncostatin-M (OSM)-11 (DOS)-like region, a Δ serrate lag-2 region (DSL) and a module at the N-terminus of Notch ligands (MNNL), with an N-terminus. Adapted from Kovall et al (123), permission from The Society of Developmental biology.

The activation of Notch-1 is initiated through the binding of its ligand Jag-1 (117, 124)

as shown in Figure 1.8. Activation of Notch-1 has been achieved using a

recombination activation gene-1 (RAG-1) mouse model (115). This model resulted in

a 15-fold increase in KSL HSCs and inhibition of HSC differentiation towards the B cell

lineage. Interestingly T cell differentiation was not impaired, providing evidence that

Notch-1 is required for HSC and progenitor fate. Walker et al (113) also found that

Jag-1 activation of Notch-1 resulted in reduced proliferation of HSCs and an increase

in G0/G1 cycling cells in vitro, therefore, Notch signalling potentially facilitates HSC

quiescence.


Figure 1.8: The activation of Notch-1 by Jag-1: A schematic diagram illustrating the activation of Notch-1 through the binding of Jag-1 in a neighbouring cell. 1. Jag-1 binds to the EGF-like proteins in the extracellular region of Notch-1, causing a conformational change. 2. A disintegrin and metalloproteinases (ADAMs) cleave sites at the intracellular Notch-1 domain resulting in a membrane tethered intracellular Notch-1. 3. γ secretases cleave the membrane tethered intracellular Notch-1. 4. Jag-1 and extracellular Notch-1 detach from the cell and intracellular Notch-1 is released. 5. Intracellular Notch-1 translocates to the nucleus and binds to the DNA binding protein CBF1/RBPjk/Su9H) Lag-1 (CSL), resulting in the recruitment of mastermind and MED8, which induce up-stream target genes. These include hairy enhancer of split (HES), hairy-related (HRT) and hairy/ enhancer of split related with YRPW motif protein 1 (Hey1). Adapted from Kopan et al (125), permission from Cell Press 30.07.14.

In addition, Karanu et al (114) reported an increase in stem cell self-renewal in the

BM of mice that had received Jag-1 activated transplanted HSCs cells compared to


controls. This is further supported by Calvi et al (82) who demonstrated an increase

in osteoblastogenesis using PPR, which resulted in an increase in Jag-1 expression as

well as an increase in HSCs in the BM. In addition, Notch-1/ Jag-1 are essential

molecules, as demonstrated using a Notch-1 KO mouse model (121). Embryos in this

model died before birth suffering from necrotic tissue, cell death and developmental

retardation from 8.5 days post-coitum (PC). Therefore, Notch-1 KO is a lethal

mutation, potentially essential for long-term definitive haematopoiesis (119). To

investigate the effects of Notch-1 in post embryonic development, a Notch-1

inducible KO model was developed which knocked-out Notch-1 after administration

of polyl-polyc (126). This resulted in a decrease in mature thymocytes, mature T cells

and an increase in cells of the B cell lineage in the thymus. This therefore, highlights

that Notch-1 signalling is required for the differentiation of cells in the T cell lineage

and when Notch-1 is deficient, this process is diminished. In addition to this, these

data also imply that Notch-1 signalling may be responsible for inhibition of B cell

differentiation.

1.3.3 Tie-2/Ang-1

Tie-2 is a cell surface receptor, which exists in the “Tie” family of receptors also

containing Tie-1. Tie-2 is expressed by ECs (127), HSCs (85) neutrophils (128), dorsal

root ganglion cells (129) as well as synovial lining cells and pericytes in patients with

rheumatoid arthritis (127). Its structure consists of an extracellular binding domain

and an intracellular kinase domain, as demonstrated in Figure 1.9.

Ang-1 is one of the primary ligands for Tie-2, which is expressed in several cell types

such as BM stromal cells, MSCs (130), osteoblast and osteoblast progenitors (85),

pericytes (131), neutrophils and monocytes (132). It belongs to a family of four of

angiopoietins, known as Ang-1 to 4. Ang-1 and Ang-4 are both receptor agonists

whilst, Ang-2 and Ang-3 are receptor antagonists and importantly, Ang-2 specifically

blocks Tie-2 activation. The structure of Ang-1 consists of a small “super clustering

region”, a coiled-coiled domain and a fibrinogen-like Tie-2 binding region (133) as

illustrated in Figure 1.10.


Figure 1.9: The structure of Tie-2. A diagram illustrating the structure of Tie-2, which consists of an extracellular region containing Ig 1-3 domains, EGF-like repeats and fibrinogen type III repeats. The intracellular domain comprises of split tyrosine kinases. Adapted from Huang et al (133), permission from Nature Reviews 30.07.14.

Figure 1.10: The structure of Ang-1. A diagram representing the structure of Ang-1, which contains a super clustering, coiled-coiled and fibrinogen-like region. Adapted from Huang et al (133), permission from Nature Reviews 30.07.14.

The binding of Tie-2 to its ligand results in dimerisation and activation of the kinase

domain by phosphorylation (134). The activation of Tie-2 by Ang-1 is essential for

several functions, which relate primarily to angiogenesis, but also to inflammation

and haematopoiesis. For example activation of Tie-2 results in downstream signalling


of ERK, NF-κB and PI3-K pathways which can result in cell survival, proliferation,

migration, anti-inflammatory, anti-permeability and vessel sprouting/ reorganisation

responses (85, 135-139). Tie-2/ Ang-1 signalling is also essential for the survival of

functioning tissues. Knock-out of Tie-2 in murine models resulted in lethality before

birth and specifically, embryos had impaired cardiac structures and underdeveloped

vasculature (140). Similar defects were also noted in Ang-1 KO mouse embryos as

well (141).

Tie-2 is also essential for haematopoiesis and is expressed in KSL HSCs (142). When

Tie-2 was knocked-out in transgenic mouse models, adult mice experienced a

dramatic reduction in haemopoietic cell lineages, therefore, demonstrating that Tie-

2 is important for HSC differentiation and proliferation (143). In support of this, Arai

et al (142) also showed that Tie-2 was important for cell cycling in HSCs. Tie-2 positive

HSCs seemed to adhere to Ang-1 positive osteoblastic cells in an osteoblastic niche

in vivo, resulting in HSC quiescence and evasion from chemotherapy (85).

1.3.4 N-cad

N-cad is a member of the cadherin superfamily. Typically, cadherins are single pass

transmembrane, calcium dependent cell-cell adhesion glycoproteins (144). The

cadherin superfamily can be categorised into four subfamilies, which are classical,

desmosomal, proto-cadherin and cadherin-like proteins. N-cad is grouped into the

category of classical Cadherins (145).

The standard structure of a Cadherin molecule consists of an extracellular region,

transmembrane helix and a cytosolic domain, as shown in Figure 1.11. The

extracellular domain, which acts as the binding domain, contains cadherin repeat

proteins, the number of which are dependent on the subtype of classical cadherin,

which are entitled I, II and III. Type I and II consist of five cadherin repeats whereas

III contain a variable number. N-cad is a type I classical cadherin (146). The homotypic

binding of N-cad between cell junctions results in several cleavages of N-cad and

subsequent activation, as shown and described in Figure 1.12.


Figure 1.11: The structure of N-cad. A diagram representing the cytosolic and extracellular structure of N-cad. The extracellular domain contains cadherin repeats whereas the intracellular cytosolic domain contains an α and, β catenin, vinculin portion, which binds to actin in the cytoskeleton of the cell. Adapted from Angst et al 2001 (147), permission from The Journal of Cell Science 30.07.14.

N-cad possesses a number of roles throughout the body due to its expression in a

variety of tissues including the heart, the nervous system and the BM. N-cad is an

essential molecule in the body as demonstrated by N-cad KO mice which died before

birth due to heart defects and malformed somites (148).


Figure 1.12: Cleavage of N-cad upon activation. 1. N-cad binds homotypically to another N-cad molecule, resulting in cleavage of the extracellular domain by metalloproteinases. 2. The extracellular domain dissociates from the cell membrane and the intracellular domain is subsequently cleaved by γ-secretase. 3. The intracellular domain dissociates from the cell membrane resulting in the release of a soluble α, β and zinc (zn) complex used in MAPK, ERK and β-catenin signalling pathways resulting in cellular invasion and migration. Adapted from Reiss et al 2005 (149).

One potential role for N-cad in the BM is its involvement in the HSC niche, where it

may anchor HSCs to osteoblasts lining the surface of bone. However, whether N-cad

is an essential molecule for HSC maintenance in an osteoblastic niche has been highly

debated previously. Several studies reported that KO of N-cad had no effect upon

haematopoiesis (77, 150, 151) and Kiel et al and Bromberg et al were also unable to

detect N-cad in purified HSCs by polymerase chain reaction (PCR) and flow

cytometry. In contrast, in support of the importance of N-cad in a HSC osteoblastic

niche, N-cad was expressed in both human primary trabecular osteoblasts as well as

the osteoblastic cell lines MG-63 and SAOS2 (152). N-cad was also found to be

expressed in KSL HSCs (83, 84, 142), which specifically localised to N-cad positive

osteoblasts with a distinct “spindle” shape. In addition, BMPR1A deficient mice

demonstrated an increase in N-cad positive osteoblasts lining the cortical and

trabecular bone surfaces, which correlated with an increase in HSCs. This therefore,

implies that N-cad positive osteoblasts may be responsible for the self-renewal or


survival of LT-HSCs (83). In addition, knock-down (KD) of N-cad in HSCs reduced their

localisation to the endosteum in the BM in vivo and significantly reduced HSC long-

term repopulating ability in recipient mice in the BM but not in the spleen (153).

Whereas, over-expression of N-cad in HSCs, resulted in slower cycling of HSCs,

promoted HSC quiescence, and a higher number of HSCs were also found in close

proximity to the bone in vivo (154).

1.4 A potential myeloma cell niche

As described in Chapter I Sections 1.2 and 1.3, there is a large amount of evidence in

support of a HSC niche. The osteoblastic niche results in quiescence of the HSCs via

the interactions of molecules expressed on the HSCs themselves and complementary

molecules expressed by osteoblasts as shown in Figure 1.13a.

The concept of MCs existing within a niche and expressing the same molecules as

those expressed by HSCs is a moderately new theory and as of yet little research has

been conducted in order to support this hypothesis. As stated previously, Chen et al

(70) found that MCs (identified by a PKH26 dye) seemed to home preferentially to

osteoblastic niches compared to vascular or splenic niches in vivo. In addition, these

cells had a higher capacity for chemoresistance compared to MCs that were PKH26

negative, when isolated and analysed ex vivo. Several reports have also

demonstrated the importance of osteoblastic cells for prostate cancer metastasis to

bone. Specifically, Shiozawa et al (155) found that using confocal microscopy,

prostate cancer cells metastasised to osteoblastic niches in bone where they

competed with existing HSCs. Upon, treatment with PTH to increase bone, or

reduction of osteoblastic cells using the inducible Col 2.3 KO mouse model, there

were increases and decreases respectively in prostate cancer metastasis to vossicles

in vivo. In addition, treatment of naïve mice with the CXCR4 antagonist AMD3100

resulted in mobilisation of HSCs out of the BM, which increased metastatic potential

upon injection of prostate cancer cells, due to reduced competition for the


osteoblastic niche. Similarly, when mice with prostate bone metastasis were treated

with AMD3100, this caused mobilisation of prostate cancer cells from the niche.

These results are also supported by Wang et al (156) who found that prostate cancer

cells homed preferentially to the lateral surfaces of bone, which was potentially

mediated by higher numbers of osteoblastic cells on that surface compared to the

medial surface. As with Shiozawa et al (155), this interaction was also inhibited using

the CXCR4 antagonist AMD3100.

A potential myeloma cell niche is illustrated in Figure 1.13b, which shows the

expression of CXCR4, Notch-1, Tie-2 and N-cad by the MCs and the interaction of

these molecules with their complementary ligands CXCL12, Jag-1, Ang-1 and N-cad,

expressed by osteoblastic cells lining bone surfaces. These interactions may result in

the adhesion of MCs to bone surfaces and may also propagate quiescence in these

cells, providing a mechanism for MC survival in the BM.

Currently, there is some evidence, which demonstrates the expression of the HSC

niche molecules CXCR4, Notch-1 and N-cad by MCs (which is discussed in the

introductions of the individual results chapters). However, the expression of these

molecules by MCs has not been reported in the context of an osteoblastic niche in

the current literature. In order to determine the expression of these molecules in

MCs, preclinical models are typically used. The most common existing preclinical

myeloma models are described below in Section 1.5.


Figure 1.13: A diagram representing molecule interactions in the HSC osteoblastic niche and proposed MC niche. a. Molecules expressed by HSCs are CXCR4, Notch-1, Tie-2 and N-cad. The complementary niche molecules expressed by osteoblastic cells are CXCL12, Jag-1, Ang-1 and N-cad. b. A diagram representing a proposed MM osteoblastic cell niche, where MCs express the same molecules as HSCs, which interact with their complementary ligands expressed by osteoblastic cells.

a

b


1.5 Preclinical multiple myeloma models

Several models of myeloma have been developed over the past 40 years to assess

the efficacy of drugs used in the treatment of MM. Common models previously used

include immune-deficient xenograft models using human cell lines such as HS-Sultan

(157), ARH-77 (158-160) and RPMI-8226 cells (161) injected intravenously (IV) or

intraperitoneally (IP), or S6B45 cells injected subcutaneously (SC) (162) into Severe

combined immunodeficiency (SCID) mice. Multiple myeloma disease is recapitulated

in these animals (with the exception of the SC tumours), displaying characteristics of

MM which include expansion of MCs in the BM, paralysis due to spinal cord

compression and production of paraprotein. In addition, primary patient samples

have previously been injected into human foetal long bone chips and implanted into

immunocompromised mice resulting in MC growth in the human bone implant in the

SCID-hu model of MM (163, 164). More recently MC cell lines including U266 and

RPMI-8226 have also been injected into NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ

(NOD/SCID γ) mice intra-tibially and U266 cells have been injected IV (165). Also

NOD.Cg-Rag1tm1Mom IL2rgtm1Wjl/SzJ mice, a radiation resistant alternative to

NOD/SCID γ mice (166) have been used for recapitulation of MM by injection of U266

and H929 cell lines as well as primary MCs, injected IV.

Alternatively, immune-competent myeloma models are used. These include the 5T

series of MM cells, where C57BL/KaLwRij mice are injected with murine “myeloma

like cells” (167). The 5T series of murine MCs originated spontaneously in 0.5% of

C57BL/KaLwRij mice older than two years, first reported by Radl et al (168). The 5T

series of murine models of myeloma are maintained by the isolation of 5T cells from

the BM of tumour bearing mice and re-injection into young syngeneic recipient mice,

where cells typically home to the BM, which is the primary location of disease

progression. The full list of 5T murine MCs and their characteristics are listed below

in Table 1.3. Most commonly used are the 5T2MM (49, 169, 170) 5T33MM (171-173)


and 5TGM1 (a derivative of the 5T33MM model produced by continual in vivo

passage of the 5T33MM cells (174)) models of myeloma (175-178).

Table: 1.3: The 5T series of murine models of MM

Adapted from Asosign et al (179), permission from The Hematology Journal, 31.7.14.

In this thesis, the 5T33 and 5TGM1 murine myeloma models were used. These

particular models are well characterised at The University of Sheffield and have the

advantage of using immune competent animals. 5T33MM and 5TGM1 cells can be

maintained in vitro (5T33MMvt and 5TGM1vt) and in addition to wild type (WT) cells;

green fluorescent protein (GFP) transduced cells (5T33-GFP and 5TGM1-GFP), have

been developed for visualisation by flow cytometry and fluorescent microscopy.

1.6 Summary

In conclusion, MM is a B cell malignancy that develops in a specialised

microenvironment in the BM. The specific mechanisms as to how MCs home,

colonise and survive in the BM are not fully understood and further research is

required in order to understand these processes.

Haemopoietic stem cells exist in BM niches whereby their proliferation and

differentiation is maintained by molecules expressed by the HSCs themselves and

complementary niche molecules expressed by constituents of the BM

microenvironment, which include osteoblastic cells. Myeloma cells may potentially

5TMM series Isotype paraprotein Osteolytic bone lesion Growth pattern 5T2 IgG2a κ Pronounced Moderate 5T7 IgG2b κ Sporadic Slow, smouldering

5T13 IgG2b κ Sporadic Moderate

5T14 IgG1 Mainly osteolytic, some sub-

lines osteosclerotic Moderate, some sub-

lines aggressive 5T21 IgD Not studied Aggressive 5T30 IgG2a κ Not studied Moderate 5T33 IgG2b κ Diffuse Aggressive 5TF IgG3 Mainly peripheral Moderate

5TGM1 IgG2b κ Pronounced Aggressive


exist within a similar niche where they express a similar repertoire of molecules as

the HSCs, which interact with their complementary ligands expressed by osteoblastic

cells located on the surface of bone. These interactions may in-turn result in MC

dormancy and subsequent survival in the BM. However, currently, little research has

been conducted to identify whether these molecules are expressed by the MCs and

whether they are important for the development of MM. Therefore, within my thesis

I have conducted further research using the 5T series of murine myeloma models to

explore whether MCs utilise the same molecular mechanisms as HSCs to acquire cell

dormancy.

1.7 Aims, hypothesis and objectives

1.7.1 Aims and hypothesis

The aim of this project was to test the overall hypothesis that “Myeloma cells exist

within an osteoblastic niche similar to that occupied by HSCs”.

Within this main hypothesis, further sub-hypotheses were explored:

1. Myeloma cells express the same repertoire of molecules as HSCs and their

complementary ligands are expressed by osteoblastic cells in vitro.

2. Myeloma cells express the same repertoire of molecules as HSCs when analysed

ex vivo and in vivo.

3. Treatment of murine primary osteoblast lineage cells (OLCs) with osteogenic

media (OGM) will result in osteoblast differentiation and differences in MC

adhesion to OLCs at different stages of differentiation.

4. The knock-down (KD) of HSC niche molecules in MCs will result in reduced MC

adhesion to primary OLCs in vitro and a subsequent reduction in tumour burden

in vivo.

1.7.2 Objectives

In order to test these hypotheses the following objectives were devised:


1. To determine the expression and presence of the HSC niche molecules CXCR4,

Notch-1, Tie-2 and N-cad by the 5T33-WT, 5T33-GFP and 5TGM1-GFP cell lines

using endpoint and real-time reverse transcription (RT)-PCR, flow cytometry and

immunofluorescence (IF) in vitro. To also establish the expression and presence

of the complementary ligands CXCL12, Jag-1, Ang-1 and N-cad by the MC3T3-E1

osteoblastic cell line and murine primary OLCs, using endpoint and real-time RT-

PCR, flow cytometry and IF in vitro.

2. To identify the presence of protein for CXCR4, Notch-1, Tie-2 and N-cad and

complementary ligands CXCL12 and Jag-1 by BM derived 5T33-GFP and 5TGM1-

GFP cells and spleen derived 5TGM1-GFP cell lines ex vivo using flow cytometry.

To also determine the presence of protein for CXCR4 and N-cad in 5TGM1-GFP

infiltrated BM sections in vivo using IHC.

3. To examine the effect of differentiating primary OLCs with OGM on alkaline

phosphatase (Alp) production, mineralisation, expression of differentiation

markers and HSC niche molecules and ligands using real-time RT-PCR, and

attachment to 5TGM1-GFP cells in vitro.

4. To Knock-down (KD) HSC niche molecules in 5TGM1-GFP cells using short hair-

pin ribonucleic acid (RNA) (shRNA), and to establish the effect of this on MC

attachment to primary OLCs in vitro and tumour burden in vivo.

Chapter II: Materials and methods Page 37

Chapter II: Materials and methods


2.1 General cell culture methods

All equipment, materials and reagents used in tissue culture are outlined in Appendix

1, Section 1.1, Table 1.1.

2.1.1 Cells

2.1.1.1 The 5T murine myeloma cell lines

The murine 5T33MM and 5TGM1 cell lines were kindly provided by Dr Karin

Vanderkerken (Vrije Universiteit Brussel, Belgium) and Prof. Babatunde Olukayode

Oyajobi (University of Texas Health Science Centre at San Antonio, USA) via Dr Claire

Edwards (The University of Oxford, UK), respectively.

2.1.1.2 MC3T3-E1 osteoblast cell line

The MC3T3-E1 osteoblast cell line was originally purchased from The American Type

Culture Collection (ATCC). Cells were previously isolated from the calvaria of C57BL/6

mice and sorted by Alp expression described by Kodama et al (180). The cell

population exist primarily as preosteoblasts however, they are able to differentiate

into mature osteoblasts using supplements such as β-glycerophosphate (BGP) and

ascorbic acid (asc), described by Kitching et al (181).

2.1.1.3 Primary osteoblast lineage cells

Primary OLCs were isolated at The University of Sheffield from C57BL/6 mouse pups

as described in Section 2.2.1.2, described previously by Bakker et al (182). Like

MC3T3-E1 cells they are able to differentiate into mature osteoblasts using BGP and

asc, explored thoroughly by Lian and Stein et al (183) using rat calvarial osteoblasts

and described by Yang et al (184) using murine primary osteoblasts.


2.1.2 Cell culture

2.1.2.1 5T murine myeloma cell line culture

All 5T murine MC lines were grown in “complete” Roswell Park Memorial Institute

medium (RPMI), containing 10% foetal bovine serum (FBS), 1%

penicillin/streptomycin (100 units/100 micrograms (µg)/millilitre (ml) (PS), 1% non-

essential amino acids (NEAA) (1X) and 1% (1 millimolar (mM)) sodium pyruvate (NaP),

pre-warmed to 37 degrees Celcius (°C). Cells were passaged every three days or when

confluent to ensure viability. At passage, the cells were split at a 1:5 dilution by

removing 3 ml of MCs in suspension and transferring them to a fresh T75 flask

containing 12 ml complete RPMI media. Cells were not used experimentally beyond

passage six. Cells were cultured at 37°C in 5% CO2.

2.1.2.2 MC3T3-E1 cell line culture

MC3T3-E1 cells were cultured in “complete” Minimum essential medium alpha

(MEMα) medium in the absence of deoxyribonucleosides and ribonucleosides, with

10% FBS and 1% PS, pre-warmed to 37°C. Cells were passaged every three to four

days or when confluent. At passage, the MEMα medium was removed and the cells

were washed twice in 10 ml phosphate buffered saline (PBS). The cells were

dissociated from the flask using 1.5 ml pre-warmed trypsin/

ethylenediaminetetraacetic acid (EDTA) (0.25%) at 37°C for 3 min (min). The

trypsin/EDTA was neutralised using 8.5 ml complete MEMα medium and the cell

suspension was transferred to a universal and centrifuged. All centrifugation was

conducted at 300 x gravitational acceleration (G) for 5 min at room temperature

unless stated otherwise. The cells were re-suspended in 10 ml complete MEMα

media and 2 ml was transferred to a T75 flask containing 10 ml MEMα media. Cells

were not used experimentally beyond passage six. Cells were cultured at 37°C in 5%

CO2.


2.1.2.3 Primary osteoblast lineage cell culture

All primary OLCs were cultured in “complete” MEMα, containing

deoxyribonucleosides and ribonucleosides (MEMα-Nuc), 10% FBS, 1% PS and 1%

fungizone anti-mycotic liquid (FZ), pre-warmed to 37°C. Primary OLCs not undergoing

differentiation were not used beyond passage six, whereas cells for differentiation

experiments were not used beyond passage two. Primary OLCs when passaged were

split as described in Section 2.1.2.2 for the MC3T3-E1 cells. Cells were cultured at

37°C in 5% CO2.

2.1.2.4 Cell counting

All cells were counted using a Neubauer haemocytometer, where 10 microlitres (µl)

of a cell suspension was combined with 10 µl trypan blue. Ten µl of the mixture was

placed under a coverslip attached to the haemocytometer and the cells were

counted as shown in Figure 2.1.

Figure 2.1: A schematic diagram of a haemocytometer: a. The dimensions and volume of each individual square. b. The volume of each 4x4 square grid. c. Cell number calculation. d. An example of cell number calculations.

2.1.2.5 Cell thawing

Each vial of cells was removed from liquid nitrogen and rapidly thawed by partial

emersion in a 37°C water bath. The thawed cells were transferred to a universal

containing 4 ml pre-warmed appropriate media, centrifuged and washed in 5 ml of

a b

c

d


media. The wash step was repeated twice to ensure removal of dimethylsulphoxide

(DMSO). After the final wash, the MCs were seeded at 1x106 cells per 12 ml of media

into a T75 flask and the MC3T3-E1 and primary OLCs were seeded at 5x105 cells per

12 ml of media into a T75 flask.

2.1.2.6 Cell freezing

Cells were counted, centrifuged and resuspended in 1 ml 10% DMSO/90% FBS per

1x106 cells. Each ml was transferred to a cryovial and added to a freezing chamber

containing 2-propanol. Cells were left over-night at -80°C, after which the vials were

transferred to a liquid nitrogen dewar. This method allowed for uniform slow freezing

of the cells and prevented ice crystal formation.

2.1.2.7 Primary osteoblast lineage cell differentiation

For all differentiation experiments, primary OLCs were seeded at a density of

6000/cm2, into a variety of different sized wells and flasks depending on the analysis.

The design of the individual aspects of the differentiation experiments are described

in the methods section in Chapter V. Three days after seeding, the cells were cultured

with or without media containing fresh BGP (10 mM) and asc (50 µg/ ml) (aliquoted

at 100X and 1000X concentrations respectively and frozen at -20°C prior to use),

referred to as OGM. When the cells were replenished, the old media was removed,

the cells were washed twice in PBS and standard MEMα-nuc or OGM was then added.

The cells were cultured in this manner every two-three days.

2.2 Primary animal cells and tissues

2.2.1 Animal cell isolation

All equipment, materials and reagents used for animal cell isolation are listed in

Appendix 1, Section 1.2, Table 1.2.


2.2.1.1 Animals

C57BL/6 and C57BL/KaLwRij mice were purchased from Harlan (Leicester, UK) and

the University of Leeds, respectively. All mice were housed in the University of

Sheffield’s Biological Services. All animal work was approved by the local ethical

committee and carried out under Home Office project licence numbers 40/2901 and

40/3462.

2.2.1.2 Primary osteoblast lineage cell isolation

Two to four day old C57BL/6 pups were sacrificed by cervical dislocation per the Code

of Practice for the Humane Killing of Animals under Schedule 1 (Scientific Procedures

Act 1986). The calvaria from each pup was dissected free of soft tissue under sterile

conditions and transferred into sterile PBS containing 1% PS. Each calvaria was

washed three times in 5 ml of sterile Hank’s balanced salt solution (HBSS), containing

1% FZ, by inversion. The calvariae were cut into small pieces and digested to release

the osteoblastic cells. Firstly, 1 ml sterile collagenase II (1 mg/ml) per calvaria was

added to the universal with the calvariae and incubated at 37°C for 15 min, inverting

the tube every 5 min. The supernatant was removed and discarded (fraction 0). The

calvariae were then digested in a further ml of collagenase II (1 ml per calvaria) for

30 min at 37°C, inverting every 10 min. The supernatant was removed, centrifuged

at 300 x g for 10 min, and re-suspended in 1 ml complete MEMα-Nuc medium and

stored at 4°C (fraction 1). The calvariae were washed twice in 1 ml PBS and further

digested in 1 ml sterile EDTA in PBS (4 mM, pH 8.0) per calvaria and incubated at 37°C

for 30 min, inverting the tube every 10 min. The supernatant was removed and

treated as with fraction 1 (fraction 2). The calvariae were then washed twice in 1 ml

HBSS followed by digestion using 1 ml collagenase II (per calvaria) and incubated at

37°C for 30 min, inverting the tubes every 10 min. The supernatant was removed and

treated as with fraction 1 and 2 (fraction 3). All the fractions from each of the pups

was pooled and divided between an equal number of T25 flasks to the number of

pups. An additional 2 ml of media was added to each of the flasks to make a total

volume of 5 ml. The cells were incubated until confluent (3-4 days), then pooled,

counted and frozen down as described in Section 2.1.2.6.


2.2.1.3 Bone marrow isolation for RNA extraction

The BM from 9-10 week old C57BL/6 mice was isolated to act as a positive control in

PCR and flow cytometry experiments for the detection of HSC niche molecules and

ligands. Mice were sacrificed under the Schedule 1 Code of Practice for the Humane

Killing of Animals (Scientific Procedures Act 1986) by IP injection of 100 µl

pentobarbitone (200 mg/ml) followed by cervical dislocation. The hind limbs were

carefully dissected free of soft tissue and the femora and tibiae were separated at

the knee joint. The proximal and distal ends of each bone were cut using a scalpel to

expose the BM. To isolate the BM, each hind limb was flushed using a sterile needle

(27 Gauge) containing 500 µl PBS and the content of each bone was transferred to a

1.5 ml Eppendorf. The samples were centrifuged, the supernatant removed and the

cell pellets were used for RNA extraction as described in Section 2.3.1 1.

2.2.1.4 Bone marrow, splenocyte and blood isolation for flow cytometry

Bone marrow, splenocytes and blood were isolated from 9-10 week old C57BLK/6

mice to act as a positive control for initial antibody optimisation experiments. In

addition, the BM and spleen were isolated from 6-8-week-old C57BL/KaLwRij mice

injected with 5T33-GFP or 5TGM1-GFP cells for the detection of HSC niche molecules

and ligands in experimental flow cytometry experiments.

Bone marrow was isolated as described in Section 2.2.1.3. However, after the

samples were centrifuged and the supernatant removed, 2 ml RBC lysis buffer (1X,

diluted in distilled water (H2O) (dH2O)), was added to each cell pellet and incubated

at room temperature for 10 min. Lysed RBCs were removed from the sample

supernatant by centrifugation and the sample was subsequently washed twice in 1

ml wash buffer (1X diluted in dH2O), provided in the RBC lysis kit. Bone marrow cells

isolated from C57BL/6 mice were counted and 1x106 cells were used per flow

cytometry sample for the detection of HSC niche molecules and ligands. Whereas,

only 100,000 BM cells, isolated from C57BL/KaLwRij mice were used per flow

cytometry sample, due to the large numbers of samples required.


To isolate the splenocytes, the spleen was carefully dissected and torn into 5 ml

Dulbecco's Modified Eagle Medium (DMEM) into a petri dish. The suspension was

filtered through a 70 µm cell strainer, to remove large pieces of tissue, followed by

centrifugation. Red blood cells were lysed and 1x106 splenocytes, isolated from the

C57BL/6 mice and 100,000 splenocytes isolated from the C57BL/KaLwRij mice were

used per flow cytometry sample.

Blood was extracted from pentobarbitone-sedated mice via cardiac puncture using a

sterile needle (27 Gauge). The blood was immediately transferred to a vacutainer

coated in EDTA to prevent coagulation and 1 ml RBC lysis buffer was added per 100

µl of extracted blood for 10 min at room temperature. Red blood cells were removed

from the supernatant following centrifugation at 600 x g for 8 min at room

temperature and the remaining cells were washed twice in RBC lysis wash buffer.

Flow cytometry was then conducted using 100,000 blood cells per sample isolated

from the C57BL/6 mice and 10,000 isolated from C57BL/KaLwRij mice.

2.2.2 In vivo models of multiple myeloma

2.2.2.1 The 5T33-GFP and 5TGM1-GFP models of multiple myeloma for detection

of HSC niche molecules and ligands

Six to eight week old C57BL/KaLwRij mice were injected IV via the tail vein with either

1x106 5T33-GFP or 2x106 5TGM1-GFP cells. All mice were sacrificed upon

presentation of hind limb paralysis (a clear indicator of tumour load). In the 5T33MM-

GFP model this occurred within 4-6 weeks post-injection (P-I), whereas in the

5TGM1-GFP model this occurred at 21-23 days P-I. Tumour burden and the presence

of protein for HSC niche molecules and ligands in the tibiae were assessed by flow

cytometry as described in Sections 2.2.1.4 and 2.4.1.2. Further details of study design

are outlined in the methods section in Chapter IV and Chapter VI.


2.2.2.2 Gene knock-down experiments in the 5TGM1-GFP model

For N-cad KD studies, a pilot experiment was firstly conducted where 11 week old

female C57BL/KaLwRij mice were injected IV with either 2x106 un-transduced

5TGM1-GFP, scrambled control cells (CTRL 25) or N-cad KD (KD 25) cells. At the first

signs of illness, the entire cohort of mice was sacrificed. Tumour burden was

determined by flow cytometry and CD138 IHC, as described in Sections 2.4.1.2 and

2.4.3.3 respectively, and bone disease was analysed using micro CT (µCT) as

described in 2.7.1.1. Further details of the study design are outlined in the methods

sections in Chapter VI.

2.3 Molecular biology methods

2.3.1 RNA extraction

All equipment, materials and reagents used for RNA extraction are outlined in


2.3.1.1 RNA methodology

To detect the gene expression of each HSC niche molecule, ligand and the plasma cell

marker CD138 from the desired tissue or cell type, the messenger RNA (mRNA) was

firstly extracted using a phenol/chloroform method (Ultraspec reagent). Myeloma

cells, MC3T3-E1 and primary OLCs were set-up in culture 24 hours prior to extraction,

as described in the method sections in the individual results chapters. After 24 hours,

the MC cells were centrifuged, washed once in 5 ml PBS and re-suspended in 1 ml

Ultraspec RNA isolation solution. Whereas, for the MC3T3-E1 cells and primary OLCs,

the media was removed, the cells were washed once in 5 ml PBS, and 1 ml Ultraspec

was added directly to each flask. In addition, the cell pellets from the murine BM

samples, following their isolation, as described in Section 2.2.1.3, were re-suspended

directly into 1 ml of Ultraspec solution. All cells were then transferred to a 1.5 ml


RNase and DNase free tube and vigorously vortexed for homogenisation for 5 min at

4°C.

To each sample, 200 µl of chloroform was added and incubated at 4°C for 5 min. This

was followed by centrifugation at 12,000 x g for 15 min at 4°C after which the content

of each tube had separated into three phases; a lower organic phase, an inter-phase

and an upper aqueous phase containing the RNA. The aqueous phase was transferred

to a new 1.5 ml RNase and DNase free tube and an equal volume of isopropanol was

added. The samples were then incubated for a further 20 min at 4°C and centrifuged

again at 12,000 x g for 5 min at 4°C to precipitate the RNA. The samples were then

washed twice in 70% ethanol diluted in diethylpyrocarbonate (DEPC) water and

centrifuged at 7,500 x g for 5 min in between washes. The RNA pellet was air dried

and solubilised in 30-50 µl of DEPC water. The samples were then stored at -80°C

until required for RT.

2.3.1.2 Quantification of RNA

The purity and concentration of all RNA samples were determined using a NanoDrop

2000 and its dedicated software. The purity of each sample was determined using

the 260/280 nanometre (nm) and 260/230 nm absorption ratios. Ribonucleic acids

absent of protein contamination (260/280 nm) had a value above 1.8 and RNA which

did not contain impurities such as phenol or ethanol (260/230 nm) had a value of

between 2.0-2.2. An example of the curves generated using the NanoDrop 2000

software are shown in Appendix 2, Section 2.1. The concentration of the RNA, which

was usually between 200-600 nanograms (ng)/µl, was also calculated using the

NanoDrop 2000 software.

2.3.1.3 RNA fragment-length analysis

Fragment length analysis of each RNA sample was established using an Agilent

Bioanalyser 2100 and analysed using 2100 Expert Software by Paul Heath at The


University of Sheffield. An RNA integrity number between 9.0 and 10 indicated

suitable RNA fragment length. Examples of fragment length analysis are shown in

Appendix 2, Section 2.2.

2.3.2 Reverse transcription

All equipment, materials and reagents used for RT are outlined in Appendix 1, Section

1.4, Table 1.4.

2.3.2.1 Reverse transcription methodology

Reverse transcription was conducted to convert mRNA into complementary DNA

(cDNA), for subsequent PCR experiments. To remove any contamination before

conducting RT, the required number of 0.5 ml RNase and DNase free tubes for the RT

reaction were irradiated for 30 min in an ultra violet (UV) hood. Two micrograms of

RNA was calculated and the required volume was transferred to an irradiated 0.5 ml

tube and DEPC H20 was added to make a final volume of 11 µl. For each RNA sample,

a superscript positive (a sample with reverse transcriptase enzyme) (RT+) and

superscript negative (a sample without reverse transcriptase enzyme) (RT-) replicate

was prepared, to control for genomic contamination.

Reverse transcription was conducted by combining random primers,

deoxynucleotide triphosphates dNTPs and DEPC H20 together in a 0.5 ml tube. The

volumes required of each reagent for one sample is shown in Table 2.1. Two

microlitres of this mixed was added to each tube containing the RNA and DEPC H20

to make a 13 µl solution. The samples were then placed in a MJ Research PTC 200

thermal cycler for 5 min at 65°C to allow for the binding of the random primers to

the RNA template. After 5 min, the samples were removed from the thermal cycler

and cooled at 4°C for at least 1 min.


Table 2.1: Reagents required for one sample in the first phase of reverse transcription

This was followed by combining first strand buffer, dithiothreitol (DTT), RNase

inhibitor (RNasin) together and then adding either SuperScript RT enzyme (RT+) or

DEPC H20 (RT-). The volumes added are shown in Table 2.2.

Table 2.2: Reagents required to make one RT+ and RT- sample in the second phase of reverse transcription

The entire 7 µl of the mix was then added to each RNA sample and placed in the

thermal cycler. The RT samples were heated at 25 °C for 5 min, followed by 50 °C for

55 min, 70 °C for 15 min and 4 °C thereafter, as described in Table 2.3. The cDNA

samples were then removed from the thermal cycler, diluted approximately to 20

ng/µl, using nuclease free water and stored at -20°C until required for PCR.

Reagent Stock solution Working solution Volume

(µL)

DEPC H2O n/a n/a 0.5 Random primers 500 ng/ µl 250 ng 0.5

dNTPs 10 nM 5 nM 1

Reagent Stock solution Working solution RT-

volume (µL)

RT+ Volume

(µL)

First strand buffer 5X 1X 4 4 DTT 0.1 M 5 mM 1 1

RNasin 1 unit/ µl 1 unit 1 1

SuperScript III, Reverse transcriptase

enzyme 200 units/ µl 200 units 0 1

DEPC H2O n/a n/a 1 0


Table 2.3: Temperatures and times used for reverse transcription

Temperature (°C) Time (Min)

25 5 50 55 70 15 4 Thereafter

2.3.3 Endpoint PCR

All equipment, materials and reagents used for endpoint PCR are outlined in


2.3.3.1 Principals of endpoint PCR

Endpoint PCR is a tool used for the amplification of a sequence of DNA from a

particular gene of interest (GOI). Once amplified, the gene product can be visualised

by agarose gel electrophoresis. The basic principle of PCR, which is displayed in Figure

2.2, involves the use of a heat stable enzyme known as Taq polymerase, isolated from

the bacterium Thermus Aquaticus. Within the PCR reaction, specific primers bind to

their complementary nucleotides on the cDNA template, initiating the binding of Taq

polymerase and subsequent transcription of a complementary strand of the

template by the binding of dNTPs to their complementary nucleotides. After a

number of cycles, the desired product is amplified sufficiently for visualisation in the

agarose gel.


Figure 2.2: A schematic diagram displaying the process of gene amplification using PCR. 1. DNA is denatured to form two complementary strands. 2. Primers then anneal to their complementary sequence on each strand of DNA, which initiates binding of Taq polymerase. 3. Taq polymerase then transcribes the complementary strand by addition of dNTPs and elongation of the strand. This cycle is repeated to amplify the gene of interest (GOI).

2.3.3.2 Endpoint PCR methodology

An endpoint PCR mix was made by combining magnesium chloride (MgCl2),

ammonium buffer, dNTPs, sense and anti-sense primers, Taq polymerase and RT-PCR

H2O. Forty-eight microlitres of the PCR mix was added to 2 µl of cDNA (approximately

40 ng) to make a final volume of 50 µl. The volume required of each reagent for one

sample is described in Table 2.4.


Table 2.4: Reagents required for one endpoint PCR sample

Samples were loaded into the thermal cycler. The standard settings for the PCR

reactions were as follows; denaturation at 95⁰C for 5 min, annealing for 1 min

(temperatures varied for each set of primers, see Appendix 2, Section 2.4, Table 2)

and elongation at 72⁰C for 1 min. These steps were then repeated for the number of

cycles required; see Appendix 2, Section 2.4, Table 2 for cycle numbers. This was

followed by a final elongation step at 72⁰C for 10 min and the samples were then

cooled to 4⁰C. The end PCR product was either stored at 4°C short-term or at -20°C

long-term.

2.3.3.3 Primer design

Primers were designed using the following guidelines: 20 base pair (BP) in length,

G+C content between 40%-60%, no repeated lengths greater than 3 BP, no repeat

dinucleotides e.g. ATATATATA, primers must not be complementary to avoid primer

dimer, must cross exon boundaries, the melting temperatures (Tm) must not differ

by >5°C between the primers, the 3’ end should contain a G or C (G/C Clamp).

The full sequence of each set of primers and their Tms are shown in Appendix 2,

Section 2.5, Table 2.2.

Reagent Stock solution Working solution

Volume (µL)

MgCl2 50 mM 3 mM 3 Ammonium

buffer 10X 1X 5

dNTPS 10 nM 1 nM 5 Sense and anti-sense primers 5 µm 0.2 µm 2

Taq polymerase 5 units/ µ 2.5 units 0.5

RT-PCR H20 n/a n/a 30.5


2.3.4 Gel electrophoresis

All equipment, materials and reagents used for gel electrophoresis are outlined in


2.3.4.1 Principals of gel electrophoresis

Gel electrophoresis allows for the separation and subsequent visualisation of PCR

products. The amplified DNA is loaded into an agarose gel and when a current is

applied, the negatively charged DNA moves through the gel towards the anode. The

larger the size of DNA product, the slower it moves through the gel due to the

restriction of the agarose pore size. Therefore, larger products are visualised nearer

to the top of the gel. A DNA ladder is also used to estimate the BP size of the DNA

product. Ethidium bromide in the agarose binds to DNA present in the sample and

subsequently fluoresces upon exposure to UV light. Therefore, gene expression of

the sequences amplified by PCR are visualised as a white band.

2.3.4.2 Gel electrophoresis methodology

The agarose gel for the electrophoresis was made using trizma base, acetic acid, EDTA

(TAE) buffer (pH 8.0) and agarose. The quantities of each reagent required to make

one litre of 50X TAE buffer demonstrated in Table 2.5.

Table 2.5: Reagents and the quantities required to make 1 litre of 50X trizma base, acetic acid and EDTA buffer

Reagent Stock solution Working solution Amount required

Trizma base n/a 2 M 242 g EDTA pH 8.0 0.5 M 50 mM 100 ml

Glacial acetic acid Glacial 1 M 57.1 ml dH2O n/a n/a 842.9 ml

The 50X solution was then diluted 1:50 using dH2O before use, to produce a solution

containing 40 mM tris base, 20 mM glacial acetic acid and 1 mM EDTA, pH 8.0. To

make the gel, agarose was added to the TAE buffer at a 1.6% concentration and then


dissolved by heating in a microwave. The agarose was then cooled to approximately

60°C and ethidium bromide was added at a final concentration of 0.5 µg/ ml. The

agarose was then poured into a gel mould and electrophoresis combs were added to

form wells. The gel was then left to cool to room temperature for 45 min. Once

cooled the gel was transferred to an electrophoresis tank containing 1X TAE buffer,

which covered the surface of the gel. The combs were removed to allow for the

addition of PCR products into the wells. Twenty microlitres of PCR product and 5 µl

of loading buffer (1X) were added to an irradiated 0.5 ml RNase and DNase free tube.

Fifteen microlitres of RT+ and RT- samples were loaded into neighbouring wells

preceded by 15 µl of RT-PCR H2O and 15 µl of the hyperladder IV. The DNA was then

separated at 120 Volts using a power pack for 40 min and visualised using a Bio-rad

Gel Doc XR+ Imager system and Quantity One 4.6.8 software.

2.3.4.3 Gene sequencing

To confirm the specificity of the PCR products, each PCR sample was sequenced in-

house by the University of Sheffield Medical School Genomics Facility. After

sequencing, the exact homology was analysed by comparing the sequencing results

with the expected template and GenBank archives.

2.3.5 Real-time PCR

All equipment, materials and reagents used for real-time PCR are outlined in


2.3.5.1 Principals of real-time PCR

Real-time PCR uses the basic principles of PCR however, rather than displaying the

results in an agarose gel the results are assessed quantitatively in “real-time”. In

addition, rather than using primers, intercalating dyes or fluorescent hydrolysis

probes are used. The intercalating dyes bind directly to the DNA whereas, the

fluorescent probes, bind to complementary nucleotides on the cDNA template.


Fluorescent probes are made of DNA oligonucleotides with a 5 prime end (5’) bound

to a reporter molecule (such as 6-carboxyfluorescein (FAM)) and a 3 prime end (3’)

bound to a quencher molecule, such as minor groove binder (MGB), Their use in real-

time PCR is shown in Figure 2.3.

Figure 2.3: A schematic diagram displaying binding of a probe to its DNA target sequence and sequential fluorescence emission. The probe consists of a DNA sequence complementary to its target gene, in addition to a reporter (R) and quencher (Q) molecule. a. The probe binds to the complementary nucleotides on the cDNA template. At this point, when the molecules are in close proximity, the fluorescent out-put is low. b. The reporter is cleaved by Taq polymerase. c. Separation of the reporter and quencher molecules result in an increase in fluorescent out-put. With every cycle the fluorescence increases producing a cycle threshold (Ct) value, which indicates the number of cycles required to cross the set threshold. The greater the abundance of target sequence, the greater the fluorescence and fewer cycles required to cross the threshold, resulting in a lower Ct value.

2.3.5.2 Real-time PCR methodology

To remove any contamination, a 384 well plate, 0.5 ml RNase and DNase free tubes,

RT-PCR H20, a microseal adhesive cover and tips were placed in the UV hood for 30

min. For each reaction, a mix containing RT-PCR H20, Master Mix and probes was

made using the volumes for one sample demonstrated in Table 2.6.

a

b

c


Table 2.6: Reagents and the quantities required for one real-time PCR sample

Reagent Stock solution Working solution Volume (µl)

RT-PCR H2O n/a n/a 2.5 Master Mix 5X 1X 5

Probe 0.5

The entire 8 µl mix was transferred into a well of the 384 well plate and to this, 2 µl

of cDNA (approximately 40 ng) was added. The plate was covered using the microseal

adhesive cover and centrifuged at 300 x g for 1 min at room temperature to ensure

all liquid was combined at the bottom of the well. The probes used were designed to

bind to complementary nucleotides on the cDNA template for the GOI and

housekeeping (HK) genes, which included β2M, hypoxanthine-guanine

phosphoribosyltransferase (HPRT) or transferrin receptor (TFRC) designed by Life

Technologies. The list of probes used and their reference numbers are in Appendix 2,

Section 2.6, Table 2.4. The plate was analysed using the HT 7900 real-time PCR

machine and SDS 2.3 software. Cycle conditions, which were standard for all real-

time PCR machines per Applied Biosystems, are described in Figure 2.4.

Figure 2.4: A diagram describing the conditions used for real-time PCR. At stage one the reaction mixture containing the probes, Master Mix and cDNA was heated to 50°C for 2 min to activate all reaction reagents, followed by specific Taq activation for 10 min at 95°C, at stage 2. Stage 3 and 4 conditions were repeated for 40 cycles whereby the cDNA and complementary strands were denatured at 95°C for 0.15 min followed by annealing of probes and elongation at 60°C for 1 min.

Gene expression of each molecule was either presented using their Ct values or 1/ΔCt

value (calculated by normalisation to a HK gene), or using the fold change or

percentage of gene expression using the ΔΔCt method. The latter is shown in greater


detail in Table 2.19. Genes which were not expressed and had a Ct value of greater

than 35 were referred to as undetermined (UD).

2.3.5.3 Probe efficiency

Probe efficiency experiments were conducted before use of the probes, the results

for which are shown in Appendix 2, Section 2.6, Figure 2.3. Serial dilutions of control

murine BM cDNA (101-104), were added as in the reactions described above. The data

was presented graphically using the Ct values and trend lines were plotted. Primer

efficiency was calculated using this formula: Efficiencyx = 10(-1/slope) -1 * 100. The

percentage efficiency values for each of the probes is shown in Appendix 2, Section

2.6, Table 2.6. All probes were within the acceptable range of efficiency of between

90-110% as advised by the manufacturer, with the exception of TFRC, which was not

used in future experiments.

2.4 Cell biology

2.4.1 Flow cytometry

All equipment, materials and reagents used for flow cytometry are outlined in


2.4.1.1 General principles of flow cytometry

Flow cytometry is a technique which allows for the quantification of protein on a per

cell basis, as illustrated in Figure 2.5. In summary, a fluorescently labelled antibody

binds to its antigen if present on the cell. The cell with the bound fluorescent

antibody or a cell absent of fluorescent antibody moves through the pressurised

system passing through beams of light and subsequent detectors, which detect the

forward scatter (cell size) and side scatter (granularity) and therefore providing


information with regards to the general structure of the cell. In addition, a series of

lasers excite the fluorochromes attached to the bound antibodies and the light

reflected passes through dichromic mirrors to the detectors, which detect specific

wavelengths dependent on the fluorochrome used. The results are then analysed

using specialised computer software. If fluorescence is detected by the software, this

implies that an antibody is bound, and the software expresses the results as a

percentage of the cells with bound antibody. Therefore, this provides information

about the percentage of the cell population, which are positive for the antigen of

interest.

Figure 2.5: A schematic diagram illustrating the principles of flow cytometry a. The cells are incubated with a fluorescent antibody. b. Upon flow cytometry, the cells travel through the flow cytometry machine via the flow cell and the antibodies are excited by a series of lasers. c. Their emission wavelengths are reflected by dichroic mirrors and pass through various filters (FL-1 to FL-4, forward scatter (FSC) and side scatter (SSC)). d. The emission signal detected is analysed and interpreted by computer software. Adapted from Brown et al (185) (2000), permission from The Journal of Clinical Chemistry.


2.4.1.2 Flow cytometry methodology

The murine BM, splenocytes and blood samples were prepared for flow cytometry

as described in Section 2.2.1.4. All cell lines were counted and set-up in tissue culture,

24 hours prior to flow cytometry as described for RNA extractions in Section 2.3.1 1.

On the day of flow cytometry, the MCs were harvested directly from the 6 well plate,

transferred to a universal, centrifuged and washed twice in 10 ml PBS. The media

from the MC3T3-E1 cells and primary OLC samples in the T75 flasks was removed and

the cells were washed twice in 10 ml PBS before 5 ml cell dissociation solution was

added. Flasks were then incubated at 37⁰C for 3 min to gently detach the cells from

the surface of the flask without impairing surface protein expression. The remaining

cells were gently scraped from the surface using a cell scraper and 10 ml of

appropriate medium was added to the harvested cells. The cells were centrifuged

and re-suspended in 10 ml of media, which was transferred to a T25 flask. This

suspension of cells was then left standing upright in the incubator at 37⁰C for 3 hours

to restore any receptor activity that may have been impaired. After the allocated

time, the cells were harvested and washed twice in 10 ml PBS.

All cell types (which included the BM, splenocytes, blood cells, MCs and osteoblastic

cells) were centrifuged and 200 µl 10% goat serum diluted in 0.5% bovine serum

albumin (BSA) in PBS was added per sample, transferred to a well in a V-bottomed

96 well plate and incubated for 20 min at 4°C to block Fc receptors. After which, the

cells were centrifuged as normal for the myeloma, osteoblastic, BM and spleen

derived cells or 600 x g for 8 min at room temperature for the blood. They were then

re-suspended in 100 µl 0.5% BSA and surface staining of CXCR4, Jag-1, Tie-2 and

CD138 was conducted by adding an optimum concentration of antibody and matched

dose of isotype control antibody (shown in Table 2.7) to each sample re-suspended

in 100 µl 0.5% BSA. Samples were incubated at 4⁰C for 30 min in the dark. After

incubation, the samples were washed three times in 200 µl 0.5% BSA, using the

centrifugation settings described previously. After the final wash, the samples were


re-suspended in 300 µl 0.5% BSA, passed through a 70 µM cell strainer and

transferred to flow cytometry tubes.

Notch-1 and CXCL12 protein detection required an intracellular staining technique,

whereby the cells were centrifuged after blocking and fixed in 100 µl 4%

paraformaldehyde (PFA) per sample, provided in the permeablisation kit, for 10 min

at room temperature. After fixation, the samples were centrifuged as before, washed

once in 200 µl PBS and then washed twice in 200 µl permeablisation buffer, provided

in the permeablisation kit (diluted to 1X using dH2O). The cells were then re-

suspended in 100 µl permeablisation buffer, the optimum antibody concentration

was added to the cells and the samples were incubated at 4⁰C for 30 min in the dark.

After incubation, the cells were washed three times in 200 µl permeablisation buffer,

re-suspended in 300 µl 0.5% BSA and the cells were passed through a 70 µm strainer

and transferred to flow cytometry tubes.

Table 2.7: Antibodies used in flow cytometry

Antibody specificity

Host Isotype Fluorochrome Amount of antibody

Supplier

Anti-CXCR4 Rat IgG2b ĸ APC 0.5 µg eBioscience CXCR4 isotype Rat IgG2b ĸ APC 0.5 µg eBioscience

Anti-CXCL12 Mouse IgG1 APC 0.25 µg R&D Systems CXCL12 isotype Mouse IgG1 APC 0.25 µg R&D Systems

Anti-Notch-1 Mouse IgG1 ĸ PE 1.0 µg eBioscience Notch-1 isotype Mouse IgG1 ĸ PE 1.0 µg R&D Systems

Anti-Jag-1 Armenian hamster

IgG PE 0.5 µg eBioscience

Jag-1 isotype Armenian hamster

IgG PE 0.5 µg Cambridge Bioscience

Anti-Tie-2 Rat IgG1 ĸ PE 0.5 µg eBioscience

Tie-2 isotype Rat IgG1 ĸ PE 0.5 µg Cambridge Bioscience

Anti-N-cad Mouse IgG1 Dylight 650 0.25-1.0 µg Novus

Biologicals N-cad isotype Mouse IgG1 APC 0.25-1.0 µg R&D Systems

Anti-CD138 Rat IgG2a ĸ APC 0.5 µg BD

Pharmingen CD138 isotype Rat IgG2a ĸ APC 0.5 µg eBioscience


To detect the viability of cells, where possible, cells were stained with 2 µl of either

TO-PRO 3 iodide (TO-PRO-3) (final concentration of 2 nM) or propidium iodide (PI)

(final concentration of 50 µg/ml). The samples were analysed using a Fluorescence-

Activated Cell Sorting (FACS) Calibur machine (Becton Dickinson) (BD). The laser and

detectors used for analysis were dependent on the absorbance and emission spectra

of each fluorochrome. The lasers used to excite each fluorochrome and the detector

used to analyse their emissions are displayed in Table 2.8. In order to prevent spectral

overlap between the GFP in FL-1 and the PI or PE in FL-2, compensation was applied.

Compensation used to prevent FL-1/FL-2 overlap was 20%, established using naïve

BM or WT myeloma cell samples by comparing it to GFP positive samples.

Compensation to prevent FL-2/FL-1 overlap was 4.5%, established using PI and PE

negative BM or cell samples, which were compared to PI and PE positive samples.

These settings were consistently used throughout all experiments with GFP

expressing cells and PE and PI fluorochromes.

The percentage of cells which were positive for each fluorochrome and therefore,

positive for tumour burden or each molecule in 10,000-100,000 events was then

calculated using the Cell Quest software. These were calculated by drawing a gate

around the GFP or fluorochrome positive population based on naïve BM or isotype

control samples, to ensure that less than one percent of staining was evident in the

gated region. The settings for the gates then remained the same throughout each

experiment.

The Cell Quest software can also analyse mean fluorescent intensity (MFI), however,

the main focus within these experiments was the number of positive cells in the

population rather than intensity on a per cell basis, therefore, MFI was not

calculated.


Table 2.8: Laser, absorbance, emission and detector information for each fluorochrome used in flow cytometry

2.4.2 Immunofluoresence staining

All equipment, materials and reagents used for IF are outlined in Appendix 1, Section

1.9, Table 1.9.

2.4.2.1 Principles of immunofluorescence staining

Immunofluorescence was used for the visualisation and quantification of the

presence of protein by the 5T33-GFP and 5TGM1-GFP MCs. The general principals

are similar to flow cytometry whereby a fluorescent antibody binds to the antigen of

interest on cells however; the cells are bound to a fixed surface rather than in

suspension. When the cells are placed under a fluorescent microscope, bound

fluorescent antibodies attached to the cells, are excited by a light source, and light

emitted passes through a series of filters and detectors to produce an image using

the computer software.

2.4.2.2 Immunofluorescence staining methodology

The 5T33-GFP and 5TGM1-GFP cells were grown in culture as described in 2.1.2.1 and

subsequently stained for the HSC niche molecules, ligands and CD138 by IF. On the

day of IF staining, 1x105 cells per sample were transferred to a universal and washed

once in 5 ml PBS before fixation in 500 µl 4% PFA for 10 min at room temperature.

After fixation, the cells were washed twice in 5 ml PBS and re-suspended in 200 µl

Laser

excitation (nm)

Absorbance Max (nm)

Emission Max (nm)

Detector

GFP 488 490 515 FL-1 Phycoerythrin

(PE) 488 496 578 FL-2

Allophycocyanin (APC)

633 650 660 FL-4

PI 488 540 620 FL-2 TO-PRO-3 633 640 657 FL-4


PBS. The cells (1x105) were then cytospun at 500 rotations per min (RPM) for 5 min

on a medium acceleration on to a super-frost slide and then left to dry for 5 min at

room temperature before beginning the staining the procedure.

The cells were isolated using a hydrophobic barrier pen to contain staining solutions

and the slides were placed in PBS. They were then added to a humidified immuno-

tray where they were blocked in approximately 200 µl 10% goat serum in PBS for 20

min at room temperature. For cell surface staining of CXCR4, Notch-1 and CD138, the

block was “tapped” off and approximately 200 µl primary fluorescent antibody or

matched dose isotype antibody diluted in PBS was added. The dilution of each

antibody is displayed in Table 2.9. The antibodies were incubated for 45 min at room

temperature in the dark, followed by washing in PBS three times for 5 min.

For intracellular staining of CXCL12, the cells were permeablised after blocking by

washing the slides twice in a bath containing 0.5% tween diluted in PBS (PBST) for 5

min per wash. After which the slides were stained with the appropriate dilution of

antibody in 0.5% PBST for 45 min at room temperature in the dark. The slides were

then washed three times in a bath containing 0.5% in PBST, 5 min per wash.


Table 2.9: Antibodies used for immunofluorescence staining

After extracellular and intracellular staining and subsequent washing, the slides were

cover-slipped using the aqueous Prolong Gold mount containing 4',6-Diamidino-2-

phenylindole, dihydrochloride (DAPI) and 22 x 22 millimetre (mm) cover slips. The

edges were then sealed with nail varnish and the slides were left to dry for an hour

before examining by fluorescent microscopy.

2.4.2.3 Fluorescent microscopy

Immunofluorescence staining was visualised using a Leica AF6000 microscope and

LAS-AF software. The settings forvisualising both the DAPI and APC are displayed in

Table 2.10. The settings differed slightly depending on the molecule being analysed.


Host

Isotype Fluorochrome Concentration Dilution Supplier

Anti-CXCR4

Rat IgG2b ĸ APC 0.2 mg/ml 1:300 eBioscience

CXCR4 isotype

Rat IgG2b ĸ APC 0.2 mg/ml 1:300 eBioscience

Anti-CXCL12

Mouse IgG1 APC 20 µg/ml 1:5 R&D Systems

CXCL12 isotype

Mouse IgG1 APC 10 µg/ml 1:2.5 R&D Systems

Anti-Notch-1

Mouse IgG2a ĸ APC 0.2 mg/ml 1:20 eBioscience

Notch-1 isotype

Mouse IgG2a ĸ APC 0.2 mg/ml 1:20 eBioscience

Anti-CD138

Rat IgG2a ĸ APC 0.2 mg/ml 1:300 BD Pharmingen

CD138 isotype

Rat IgG2a ĸ APC 0.2 mg/ml 1:300 eBioscience


Table 2.10: Leica AF6000 microscope and LAS-AF settings used to visualise immunofluorescence staining of the HSC niche molecules, ligands and CD138 in the 5T33-GFP and 5TGM1-GFP cells

Molecule Fluorochrome Filter cube

Exposure time millisecond

(ms) Gain Intensity Binning

CXCR4 DAPI A4 250 3.1 4 1x1 APC Y5 590 10 5 1x1

CXCL12 DAPI A4 300 3.1 4 1x1 APC Y5 590 10 5 1x1

Notch-1 DAPI A4 220 3.1 4 2x2 APC Y5 480 10 5 2x2

CD138 DAPI A4 220 3.1 4 2x2 APC Y5 480 10 5 2x2

After visualisation, the pictures were exported as tagged image file format files

(TIFFs) and imported into Volocity software (Perkin Elmer) for image analysis.

Volocity software was used to manually spot count the number of DAPI stained nuclei

and APC labelled cells to calculate the percentage of cells positive for each molecule

of interest. An example of this is shown in Figure 2.6.


Figure 2.6: Quantification of immunofluorescence using Volocity software. The 5T33-GFP cells were stained with an anti-CXCR4 antibody and DAPI. Staining was visualised and analysed using Volocity software. This was conducted by manually spot counting DAPI and APC positive cells to calculate the number of cells positive for each molecule of interest. a. Manual spot counting of DAPI stained nuclei (blue) using 5T33-GFP cells. Numbers represent each DAPI positive cell. b. Manual spot counting of the CXCR4 staining (red) using 5T33-GFP cells. Numbers represent each DAPI positive cell.

2.4.3 IHC staining

All equipment, materials and reagents used for IHC are outlined in Appendix 1,

Section 1.10, Table 1.10.

2.4.3.1 Principles of IHC

Immunohistochemical techniques were developed to visualise the expression of the

HSC molecules of interest ex vivo. A “sandwich” IHC system was optimised to

visualise the presence of protein for N-cad, CXCR4 and CD138 in BM histological

a

b


sections taken from mice injected with 5TGM1-GFP cells. The basic IHC method used

is displayed in Figure 2.7.

Figure 2.7: A schematic diagram demonstrating a sandwich IHC technique. In IHC, a sandwich antibody technique can be used to detect proteins of interest. To do this a primary antibody binds to the antigen of interest. This is followed by the binding of a biotinylated secondary to its complementary antibody species. The biotin on the secondary antibody then binds to the streptavidin horseradish peroxide (SA-HRP), which binds to the peroxidase substrate 3'-Diaminobenzidine (DAB). This then produces a brown colour, which can be visualised using light microscopy.

2.4.3.2 Preparation of bones for IHC

Tibiae were dissected free of soft tissue from C57BL/KaLwRij mice which had been

injected with 5TGM1-GFP cells after 10, 14, 17 and 21 days P-I. The bones were fixed

in 4% PFA for 24 hours and then decalcified in 0.4 M EDTA pH 8.0 containing 0.5%

PFA for 2 weeks, changing the solution every other day. The bones were placed into

tissue processor cassettes and washed three times, an hour per wash, in PBS to rinse

away residual EDTA. The bones were placed into the Leica TP2010 processor to

dehydrate the tissues and infiltrate them with wax using the programme as described

in Table 2.11.


Table 2.11: Processing schedule for bones using the Leica TP2010 processor

After processing, the bones were embedded into wax in a specific orientation, which

was the same for each tibia. The wax blocks were trimmed using a Leica microtome

to expose the length of the BM (approximately 900 µm) and 3 µm sections were cut

in serial, and put on to a 45°C water bath to “float-out” for 30 min. They were then

attached to super-frost positively charged slides and placed onto a hot plate at 45°C

to remove residual creases. Finally, they were placed in an oven at 37°C to fully

adhere to the slides over-night before they were used for IHC.

2.4.3.3 IHC methodology for CD138

On the day of staining, sections were deparaffinised twice in xylene for 4 min

followed by rehydration through a series of alcohols (99%, 99%, 95% and 70%

industrial methylated spirits (IMS)) for 4 min each. The sections were then washed in

tap water for 5 min and transferred to PBS for a further 5 min. Antigen retrieval (A.

Menarini kit) was performed by diluting trypsin 1:4 in the buffer provided. The trypsin

was pre-warmed for 45 min at 37°C before adding approximately 200 µl of solution

STATION SOLUTION TIME VACUUM

1 70% Ethanol 2 hours No







8 Xylene 2 hours No

9 Xylene 2 hours No

10 Wax 2 hours Yes

11 Wax 2 hours Yes


to each slide and incubating for 10 min at room temperature. The sections were then

washed twice in PBS, with agitation for 4 min per wash. To reduce non-specific

staining, the sections were blocked by adding 200 µl 10% goat serum in PBS to each

slide for 1 hour at room temperature. The block was tapped off and the slides were

then incubated with 200 µl primary or concentration matched isotype control

antibody diluted in PBS, for 1 hour at room temperature (as shown in Table 2.12).

The sections were then washed twice in PBS for 4 min with gentle agitation. To block

endogenous hydrogen peroxidase, 200 µl 3% hydrogen peroxide (H2O2) (1:10) in PBS

was added to each slide for 30 min at room temperature, followed by washing twice

in PBS with gentle agitation. Two hundred microlitres of goat anti-rat secondary

antibody, diluted 1:300, was added to the slides for 30 min at room temperature and

slides were subsequently washed twice in PBS. This was followed by incubation with

200 µl SA-HRP at a dilution of 1:300 for 30 min at room temperature followed by

washing twice in PBS. Finally, the peroxide substrate reagent was prepared by adding

1 ml of diluent to 1 drop of DAB in the Immpact DAB kit. Two hundred microlitres of

this solution was then added to the slides for 10 min at room temperature. The slides

were washed in tap water to remove excess DAB, followed by counter staining in

Gill’s haematoxylin for 20 seconds to visualise the nuclei. The slides were then left to

“blue” in running tap water for 3 min after which they were dehydrated through a

series of alcohols (70%, 95% and 99% IMS) for 10 seconds each and a final 99% IMS

for 30 seconds. This was followed by two xylene washes, the first for 1 min and the

second for a further 3 min. The slides were then cover-slipped using 22 x 22 (mm)

coverslips with Di-N-Butyl Phthalate in Xylene (DPX). Staining was visualised by

scanning the slides using an Aperio Scan Scope scanner and images were captured at

a 2X and 40X magnification using Image Scope Software.

2.4.3.4 IHC methodology for N-cad

The method for staining N-cad in the 5TGM1-GFP infiltrated bone sections by IHC

was similar to the protocol used for CD138 (Section 2.4.3.3). However, there were

some differences with regards to buffers, antigen retrieval and the antibodies used.

The wash buffer and diluent throughout the protocol was 0.1% PBST. For antigen


retrieval, the trypsin was diluted 1:1 and was used at room temperature immediately

without pre-heating, for 15 min. In addition, an anti-N-cad antibody and dose

matched isotype control antibody was used at a 1:100 dilution (as shown in Table

2.12) and incubated for 1 hour at room temperature. The secondary antibody used

was a goat anti-rabbit, at a 1:200 dilution, incubated for 30 min at room temperature.

All other details were the same as in the CD138 protocol.

2.4.3.5 IHC methodology for CXCR4

The method for staining CXCR4 in the 5TGM1-GFP infiltrated bone sections by IHC

was similar to the protocol used for CD138 IHC. However, antigen retrieval was not

required and an anti-CXCR4 and dose matched isotype control antibody was used at

a 1:50 dilution (Table 2.12). The slides were then treated in the exact same manner

as in the CD138 protocol.

Table 2.12: Antibodies used for IHC

*This was initially at a 15 mg/ml concentration, which was subsequently diluted to 1mg/ml to match the primary antibody and then further diluted 1.100 as indicated in the table.

2.4.3.6 CD138 staining quantification

To quantify CD138 staining in tumour bearing 5TGM1-GFP infiltrated bone sections,

an Aperio Scan Scope slider scanner and Image Scope software were used. An image

of an example bone section, which has been analysed, is shown in Figure 2.8. To

analyse the bones, the total area of BM space, trabecular bone and CD138 stained


Host Isotype Concentration Dilution Supplier

Anti-CD138 primary

Rat IgG2a 0.5 mg/ml 1:300 BD Pharmingen

CD138 isotype Rat IgG2a 0.5 mg/ml 1:300 BD Pharmingen

Anti-N-cad primary

Rabbit IgG 1.0 mg/ml 1:100 Pierce Antibodies

N-cad isotype Rabbit IgG *1.0 mg/ml 1:100 Dako cytomation

Anti-CXCR4 primary

Rat IgG2b 0.5 mg/ml 1:50 BD Pharmingen

CXCR4 isotype Rat IgG2b 0.5 mg/ml 1:50 BD Pharmingen


tumour colonies were calculated (mm2). As only the cellular BM space was required,

the area of trabecular bone was subtracted from the total BM area. To calculate the

percentage of tumour burden, the total area of tumour was then divided by the

cellular BM area and multiplied by 100. This technique has also been successfully

used by others at the University of Sheffield (Dr S. Lawson personal communication).

Figure 2.8: An example image using the Aperio Scan Scope slide scanner and analysis of tumour burden using Image Scope software. A representative section of a tibia (2X magnification) infiltrated with 5TGM1-GFP cells stained for CD138 (brown). Image Scope software was used to measure the tumour burden (yellow outline) the marrow space (green outline and bone (red outline).

2.4.4 Western blotting

All equipment, materials and reagents used for Western blotting (WB) are outlined

in Appendix 1, Section 1.11, Table 1.11.


2.4.4.1 Principals of Western blotting

Western blotting was used to determine the protein levels of N-cad in 5T33-GFP and

5TGM1-GFP cells, which had been knocked-down for N-cad (method described in

2.6.4). Western blotting is a technique whereby proteins are separated through an

acrylamide gel, transferred on to a nitrocellulose membrane and visualised using a

sandwich antibody reaction similar to IHC. However, rather than using a biotinylated

secondary antibody, a HRP conjugated secondary antibody and enhanced

chemoluminescence substrate (ECL) is used. Enhanced chemoluminescence

substrate consists of H2O2 and luminol, which binds to HRP on the secondary

antibody, resulting in oxidation of the luminol and emission of iridescent light.

Chemo-luminescence is visualised by standard x-ray development and the protein

appears as a band on an x-ray film.

2.4.4.2 Cell lysis

For whole cell lysis, a nonyl phenoxypolyethoxylethanol (NP-40) substitute buffer

containing tris buffer, sodium chloride (NaCl) and NP-40 was used. This was made

using the quantities in Table 2.13. The buffer was then stored at 4°C until required.

Table 2.13: Reagents and the quantities required to make 10 ml NP-40 buffer

Reagent Stock solution Working solution

Amount required

Tris hydrochloride

(pH 8.0) 1 M 50 mM 500 µl

NaCl 3 M 150 mM 500 µl NP-40 n/a n/ 100 µl dH2O n/a n/a 8.9 ml

The 5T33-GFP or 5TGM1-GFP cells were lysed by transferring 1x107 cells per sample

into a universal, followed by centrifugation and washing once in 10 ml PBS. A

protease inhibitor cocktail (PIC) (1X) was added to the NP-40 buffer at a 1:100 dilution

and the cell pellet was re-suspended in 200 µl of NP-40/PIC solution. The samples

were then water sonicated to produce ultrasound waves that mechanically disrupt


cells to further release proteins in addition to breaking down interfering cellular DNA.

The lysates were then frozen at -20°C and stored until required.

2.4.4.3 Protein quantification

Protein quantification was conducted using a bicinchoninic acid (BCA) assay. The

assay produces a purple colour in the presence of protein, which is analysed by

spectrophotometry. The colour reaction is produced by the reduction of copper (Cu)

2+ ions to Cu+ ions by peptide bonds present in proteins. The Cu+ ions interact with

the BCA, which produces a purple by-product. The intensity of the purple is measured

by the absorbance at 562 nm using a spectrophotometer and the optical density (OD)

readings of the samples are then extrapolated from the OD readings of a set of

protein standards to calculate the quantity of protein present.

2.4.4.4 Bicinchoninic acid assay methodology

Firstly, a 100 mg/ml BSA stock was made in NP-40 buffer and this was subsequently

diluted at a 1:100 dilution to 1 mg/ml using NP-40 buffer. The concentrations used

for the standards and the volume of BSA and NP-40 diluent required are shown in

Table 2.14 and those volumes were loaded into a 96 well plate.

Table 2.14: Volumes of bovine serum albumin and NP-40 required for the bovine serum albumin standards

Protein from the 5T33-GFP and 5TGM1-GFP cells was loaded into the wells at either

a 1:5 or 1:10 dilution with NP-40 buffer, making a total volume of 10 µl, as with the

Standard Concentration (µg/ml)

Volume of BSA (µl)

Volume of NP-40 (µl)

1000 10 0 800 8 2 600 6 4 400 4 6 200 2 8 100 1 9

0 0 10


Concentration

0 100 200 300 400 500 600 700 800 900 1000

0.11

0.21

0.31

0.41

0.51

StandardCurve

Linear Fit: y = A + Bx: A B R^2

STD01 (Standards: Concentration vs MeanODValue) 0.125 0.000429 0.999

protein standards. Bicinchoninic acid and copper II sulphate were combined at a 50:1

ratio respectively and 200 µl of this solution was added to the wells containing either

10 µl of BSA standard or 5T33-GFP or 5TGM1-GFP lysates. The plate was then

incubated at 37°C for 30 min to catalyse the colour change reaction and read using

the SpectraMax plate reader at 562 nm. A standard curve was plotted, as shown in

Figure 2.9.

Figure 2.9: Increasing amounts of bovine serum albumin resulted in a standard curve using a bicinchoninic acid assay. Increasing amounts of BSA (0-1000 µg/ml) were added to a 96 well plate followed by a BCA and copper II sulphate solution and incubated for 30 min at 37°C. A colour change reaction was measured using a SpectraMax plate reader at 562 nm and the curve was plotted using SoftMax Pro software.

The concentrations of the unknown sample proteins were calculated by

extrapolation of their ODs from OD and concentration values in the BSA standard

curve using the following formula:

y=a+bx

a= The intercept

b= Slope

x= Unknown concentration


y= Known OD

Rearrangement of formula= x=(y-a)/b

Using this equation the concentration of the unknown protein was calculated, as

shown in the example below using the slope and intercept values calculated from the

standard curve in Figure 2.9.

0.396 (OD)-0.125 (Intercept)/0.000429 (Slope) =631

This concentration was then multiplied by 5 or 10 to account for the dilution factors

and to provide a concentration in µg/ml.

2.4.4.5 Acrylamide gel methodology

To make the acrylamide gel, firstly, the separator and cover plates were placed in the

casting frames and added to the casting stand. The plates were then tested for

leakages before loading the gel into them.

A 7% resolving gel was made using tris buffer, acrylamide, sodium dodecyl sulphate

(SDS), ammonium persulphate (APS) and tetramethylethylenediamine (TEMED). The

gel was made using the volumes shown in Table 2.15. The resolving gel was added to

casting plates and left to set for 45 min, at room temperature.

Table 2.15: Reagents and volumes required to make one 7% resolving acrylamide gel.

A 4% stacking gel was then made using the same reagents as for the resolving gel

however; 0.5 M tris buffer pH 6.8 was used rather than 1.5 M tris buffer pH 8.0. The

gel was made using the volumes shown in Table 2.16. The stacking gel was then

Reagent Volume

dH2O 4 ml 30% Acrylamide 1.87 ml 1.5 M tris pH 8.0 2 ml

10% SDS 80 µl 10% APS 80 µl (TEMED) 8 µl


added to the top of the resolving gel, ensuring there were no bubbles and was left to

set for approximately 45 min.

Table 2.16: Reagents and volumes required to make one 4% stacking acrylamide gel

2.4.4.6 Sodium dodecyl sulphate-polyacrylamide gel electrophoresis

In initial experiments, different amounts of protein were loaded and run using sodium

dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) to determine which

protein amount was sufficient to visualise a band for N-cad in the 5TGM1-GFP cells as

shown in Appendix 2, Section 2.7, Figure 2.4. The optimum amount of protein used was

20 µg and this was used in subsequent experiments.

The desired quantity of protein (20 µg) per sample was calculated using the

concentrations determined using the BCA assay in Section 2.4.4.4. Prior to use, laemmli

buffer (4X) was added to β-mercaptoethanol at a 1:20 dilution and the laemmli buffer

was then added to the protein samples to make a total volume of 20 µl. The biotinylated

protein ladder was also diluted 1:10 using the laemmli buffer. The samples and

biotinylated marker were then denatured by heating in a hot block at 90°C for 10 min

after which they were cooled at 4°C. Before loading, the gel (in the glass plates) was

added to the electrophoresis assembly kit and inserted into an electrophoresis tank

containing 1X running buffer. The running buffer was initially made as a 10X solution

containing tris base, glycine and SDS and diluted 1:10 in dH2O upon use. The amounts

and volumes of reagents required to make 10X running buffer and the final working 1X

solution are displayed in Table 2.17.

Reagent Volume

dH2O 3 ml 30% Acrylamide 0.67 ml 0.5 M tris pH 6.8 1.25 ml

10% SDS 50 µl 10% APS 50 µl TEMED 5 µl


Table 2.17: Amounts and volumes of reagents required to make a litre of 10X running buffer and final amounts of reagent present in 1 litre of 1X running buffer

The biotinylated protein ladder (10 µl), a pre-stained protein ladder (5 µl) and the

samples (20 µl) were then loaded in the gel. Using the Biorad power pack, a current was

applied at 80 Volts to allow for the separation of the pre-stained marker through the

stacking layer followed by 150 V for approximately 1 hour and 30 min for separation of

the proteins through the resolving gel.

2.4.4.7 Protein transfer

Following loading and separation of the proteins through the acrylamide gel, the

proteins in the gel were transferred to a hybond nitrocellulose membrane. Firstly,

the nitrocellulose membrane, which was cut to the same size as the gel, was soaked

in methanol for 5 min followed by 1X transfer buffer. Ten times transfer buffer was

made using tris base, glycine and dH20, which was diluted 1:10 to 1X using d H20 and

methanol. The amounts required for the 10X transfer buffer stock solution and final

amounts present in working 1X solution are shown in Table 2.18.

Table 2.18: Amounts and volumes of reagents required to make a litre of 10X transfer buffer and final amounts of reagent present in 1 litre of 1X transfer buffer

Reagent 10X transfer buffer stock

solution

Amounts required for 10X transfer

buffer

1X transfer buffer

working solution

Tris base 250 mM 30.3 g 25 mM Glycine 1920 mM 144 g 192 mM

Methanol n/a n/a 20% dH2O n/a 1000 ml n/a

Reagent 10X running buffer stock

solution

Amounts required for 10X running

buffer

1X running buffer

working solution

Tris base 250 mM 30.3 g 25 mM Glycine 1920 mM 144 g 192 mM

SDS (20%) 1% 50 ml 0.1% dH2O n/a 950 ml n/a


Per gel, two teflon pads and four pieces of filter paper cut to the same size as the gel

were soaked in 1X transfer buffer and were then used to form a sandwich between

the gel and nitrocellulose membrane which was placed onto a transfer cassette in

the correct orientation so that the proteins would transfer on to the membrane. The

gel and membrane sandwich was placed in a transfer cassette facing towards the

anode. The cassette was placed in a tank containing transfer buffer and the tank was

surrounded by ice. The proteins were transferred onto the membrane at 50 Volts for

7 hours.

2.4.4.8 Immuno-blot

After transfer, the membrane was removed from the transfer cassette for immuno-

blotting. Firstly, the membrane was blocked in 5% skimmed milk diluted in 0.25%

PBST for 45 min at room temperature with agitation. After blocking, the membrane

was cut to separate the biotinylated protein ladder, β-actin (42 kDa) and N-cad (135

kDa) areas, using the pre-stained marker as a guide. The pieces of membrane were

placed in separate universals containing either PBST for the biotinylated marker, β-

actin antibody in PBST at a 1:3000 dilution or N-cad antibody (used previously for the

IHC staining in Section 2.4.3.5) at a dilution of 1:1000 in PBST for 1 hour, room

temperature with agitation. The membranes were then washed three times in PBST

for 5 min per wash with agitation, followed by an incubation with an anti-rabbit HRP

secondary at a dilution of 1:10000 in PBST for the β-actin and N-cad or the anti-biotin

HRP secondary antibody at 1:1000 in PBST for the biotinylated ladder for 45 min at

room temperature with agitation. The membranes were then washed three times in

PBST for 5 min per wash on a tube roller.

2.4.4.9 Immuno-blot development

After the final wash, the membranes were pieced back together onto a sheet of

plastic. Using the ECL kit, solution A was combined with an equal quantity of solution

B and this was then added to the membrane, which was then covered in plastic and

smoothed down to remove air bubbles. Chemo-luminescence was then visualised in


a dark room. The membrane in the plastic was placed inside a developing cassette

with the film. The membrane was exposed for a series of different times to optimise

the exposure. Two different ECL kits were used, a Thermo Scientific kit (30 seconds

of exposure) and a Santa Cruz kit (5 min of exposure). The film was then placed into

a developing solution until bands began to appear (approximately 30 seconds),

followed by washing in running tap water for 1 min after which the film was fixed

(fixer solution) for 2 min and washed in running tap water for 5 min. The film was

then left to air-dry over-night.

2.4.4.10 Western blotting quantification

After immuno-blotting and developing, the bands were quantified. This was

conducted by scanning the film using the Epson 4990 scanner and software using a

16-bit grey-scale. Once scanned, the bands were measured by densitometry using

ImageJ software, which provided a density value for each band present on the film.

Data was displayed as a fold change and percentage of protein KD. This was

calculated by normalising the band density of N-cad protein to the HK (β-actin) band

density, followed by comparing this value between the CTRL and N-cad known-down

samples to provide a fold change and percentage of protein KD. An example of these

calculations are shown in Table 2.19.


Table 2.19: Calculations to determine the percentage of protein knock-down between the CTRL and KD transduced cells

2.4.5 Myeloma cell growth curves

2.4.5.1 Growth curve methodology

To assess the effect of the N-cad KD upon cell proliferation in vitro, growth curve

experiments were conducted. The untreated, CTRL and KD 5TGM1-GFP cells were

seeded at a density of 5x103 cells into a 96 well plate in 200 µl complete RPMI media.

Cells were counted every 2 days and media was replenished on days 0, 2, 4, 6, 8 and

10. Cell proliferation was presented in a graphical format to visualise each phase of

growth, which included the lag, exponential, stationary and death phase as

demonstrated in Figure 2.10. The doubling-time in the exponential phase was

calculated using the following formula:

1/ ((log cell number at the end of the exponential phase- log cell number at the

beginning of the exponential phase) 3.32/ time)

Treatment β-actin N-cad Normalisation Mean of

normalisation Fold

change %

HK Protein of

interest (POI)

POI/HK Mean of all replicates

KD/CTRL 1-

KD/CTRL*100

CTRL 17.21 19.34 1.12377964

CTRL 17.50 24.13 1.37873507

CTRL 15.12 28.65 1.89420087

CTRL Mean

1.46557187 100

KD 20.03 8.19 0.40905687

KD 16.77 9.54 0.56843925

KD 13.36 10.16 0.75999102

KD mean 0.57916238 0.40 40


Figure 2.10: A typical cell growth curve. When cells expand in media over-time, they go through a series of growth phases. Firstly when seeded there is an initial phase of slow growth (lag phase) which is followed by a rapid proliferation phase (exponential phase). As nutrients and media decrease, division occurs at a steady rate and there are equal numbers of cells dying as there are dividing (plateaux phase). However, when nutrients and space become too depleted less division and an increase in cell death occur (death phase).

2.5 Primary osteoblast lineage cell differentiation

All equipment, materials and reagents used for all primary OLC differentiation

experiments are outlined in Appendix 1, Section 1.12, Table 1.12.

2.5.1 Alkaline phosphatase production analysis

2.5.1.1 Principals of alkaline phosphatase analysis

Alkaline phosphatase is a well-known marker of osteoblast differentiation and is up

regulated in the early differentiation of preosteoblasts to mature osteoblasts.

Alkaline phosphatase production by a cell can be determined using a p-nitrophenyl

phosphate (PNPP) reaction based on a colour change described by Sabakbar et al


(186). This reaction can be detected using a plate reader by spectrophotometry

(Figure 2.11).

Figure: 2.11: The chemistry of the reaction between alkaline phosphatase and p-nitrophenyl phosphate. Alkaline phosphatase hydrolyses PNPP, resulting in the formation of nitrophenolate, which produces a yellow colour. The more Alp present the quicker the reaction and colour change occurs.

2.5.1.2 Alkaline phosphatase assay methodology

Alkaline phosphatase assays were conducted on days 0, 3, 7, 10, 14, 21 and 28 post-

OGM treatment. On the day of the assay, the media was removed and the cells were

washed twice in 150 µl of ice cold PBS. The cells were lysed in 0.1% triton in PBS and

shaken on a rotating platform for 20 min. Whilst the cells were lysing, one PNPP

tablet and one tris buffer tablet were dissolved in 20 ml of dH20, forming a final

solution containing 1 mg/ml PNPP and 0.2 M tris buffer. Two hundred microlitre of

this solution was then added to the lysed cells, avoiding production of air bubbles

and the plate was immediately read using the SpectraMax plate reader. The plate

was read using the kinetic settings, every 5 min for 90 min, to provide a read out of

OD readings. The higher the intensity of the colour reaction, the higher the OD, as

shown in the example in Figure 2.12.


Figure: 2.12: Alkaline phosphatase production in primary osteoblast lineage cells cultured in standard MEMα-nuc media compared to those in osteogenic media. Primary osteoblast lineage cells were cultured with standard MEMα-nuc or OGM. After 14 days, the cells were lysed using 0.1% triton and incubated with PNPP. This resulted in a colour reaction when in the presence of Alp. Results were analysed by spectrometry using a SpectraMax plate reader, which provided OD readings over the course of 90 min. Cells that were cultured in OGM had higher OD readings, which plateaux before those cultured with the standard MEMα-nuc media.

2.5.1.3 DNA quantification methodology

After conducting the Alp experiments, the DNA of the same cells was quantified using

a PicoGreen® assay. This was conducted to enable normalisation of the Alp

production to DNA concentration. The DNA standard (100 µg/ml) in the Quant-it

PicoGreen® kit was diluted 1:50 (2.0 µg/ml) with tris/EDTA buffer. The DNA standards

were then set up in duplicate at 0, 50, 100, 200, 400, 600, 800 and 1000 ng/ml of

DNA in a 100 µl volume and transferred into a 96 well plate. Fifty microlitres of each

sample used for Alp detection was then transferred into the 96 plate and to this, 50

µl of tris/EDTA buffer was added. The PicoGreen® reagent was prepared by thawing

a 25 µl aliquot and diluting it with 9975 µl of tris/EDTA buffer. One hundred

microlitres of this solution was then added to each well. The plate was wrapped in

foil to avoid bleaching and mixed gently on a rocker for 5 min before reading on a

SpectraMax plate reader with an excitation at 488 nm and emission at 535 nm.


A standard curve was generated and the unknown DNA concentrations were

calculated as with the BCA assay described in Section 2.4.4.4. The DNA concentration

of each sample was multiplied by 4 to take into account the dilution factor and the

final concentration was expressed as ng/ml.

2.5.1.4 Calculating alkaline phosphatase activity

To calculate Alp activity, Beer-Lambert Law was used:

U= (ODt1-ODt0) x V

t X ε X l

OD t1: The OD at the end of the reaction where the reaction reaches a

plateaux

OD t0: The OD at the beginning of the linear part of the reaction

V: Volume of reaction (220 µl)

t: Reaction time (min)

ɛ: Molar coefficient of the PNPP (18.5 M-1 centimetre (cm)-1)

l: Light path (0.635 cm)

The light path, which is the length in cm in which light travels in the plate reader to

reach the reaction, was calculated as follows:

V=πR2h

Or: H= V/ πR2

V= the volume= 0.2 cm3

H= Height of the well or light path

R2= Radius of the well 0.3175 cm2= 0.1008 cm


∏= Pi= 3.142

Therefore, light path= 0.63 cm

This calculation was then normalised to the amount of DNA in µg by dividing the Alp

calculation by the amount of DNA present in the sample. This was then presented as

U/µg of DNA, a common way to present the data in recent literature (187). An

example of the full calculations for this are shown in Appendix 2, Section 2.9, Table

2.6.

2.5.2 Primary osteoblast lineage cell mineralisation analysis

2.5.2.1 Primary osteoblast lineage cell mineralisation assay methodology

Osteoblast mineralisation assays were conducted on days 7, 14, 21 and 28 post-

differentiation. Media was removed from the 24 well plates and washed twice in 1

ml PBS. The cells were fixed in 100% IMS over-night at 4°C and then washed twice in

PBS. The wells were stained with 1 ml 40 mM pH 4.2 alizarin red (a calcium binding

dye) in dH20 per well for 90 min at room temperature. Following this, the cells were

washed in 95% IMS, approximately four times to remove excess alizarin red. All of

the liquid was then removed and the cells were air-dried over-night. Plastic discs

were then added to each well and the intersections of the wells were filled with

cotton wool to allow for contrast when scanned using an Epson 4990 scanner using

the settings of: reflective, 24-bit colour and 2400 dots per inch (dpi). Mineralisation

was analysed using ImageJ software. This was conducted by opening the

mineralisation image file (TIFF) and setting it as a grey scale image. The threshold was

then adjusted so that no mineralisation nodules were excluded or over-exposed and

the threshold was then the same for each well within that plate. The software then

calculated the percentage of nodules in the well area.


2.5.3 Myeloma and primary osteoblast lineage cell co-cultures

2.5.3.1 Myeloma and primary osteoblast lineage cell co-culture methodology

To establish whether MCs adhere to primary OLCs in culture and whether the stage

of differentiation affects this adhesion, co-culture experiments were conducted.

Primary OLCs were set up and cultured with OGM as described previously in Section

2.1.2.7. On days, 3, 14 and 28 post-differentiation the media was removed and the

cells were washed twice in 150 µl PBS. Un-transduced, CTRL and KD 5TGM1-GFP cells

were seeded at a density of 5x104 into wells with or without primary OLCs, in a total

volume of 200 µl complete RPMI media per well. The plates were incubated for 1

hour and 6 hours after which, the media was removed and the wells were washed

four times in 200 µl PBS. This number of washes was sufficient to remove the majority

of cells from the wells, which did not contain primary OLCs. After the fourth wash a

final 200 µl of PBS was added to the wells and the adhesion of the 5TGM1-GFP cells

to the primary OLCs was visualised using fluorescent microscopy to quantify the

number of GFP positive cells present.

2.5.3.2 Co-culture analysis

The GFP positive MCs were visualised using the Leica AF6000 microscope and the

settings are demonstrated in Table 2.20.

Table 2.20: Settings used for the Leica AF6000 microscope

*To detect primary osteoblast lineage cells. **To detect 5TGM1-GFP cells.

Contrast/ fluorochrome

Filter cube

Exposure time

millisecond (ms)

Gain Intensity Binning

*Phase EMP

(phase) 11 2.9 32 1x1

**GFP L5 262 7.9 4 1x1


Five pictures were taken over the entire area of the well as shown in Figure 2.13 and

each image was exported as a TIFF and analysed using automated Volocity software,

which calculated the number of GFP positive cells in each image, as shown in Figure

2.14.

Figure 2.13: Analysis of adherent 5TGM1-GFP cells after 1 and 6 hours co-culture with primary osteoblast lineage cells. Fifty thousand 5TGM1-GFP cells were seeded on to primary OLCs, which had been differentiated for 3, 14 and 28 days. After 1 and 6 hours, images were taken of adherent 5TGM1 cells in each well in five different areas (0.012 cm2 per tile) using the Leica AF6000 microscope.


Figure 2.14: An example of automated green fluorescent protein positive cell counting using Volocity software. The TIFF files of 5TGM1-GFP and primary OLC co-culture tile images were loaded into Volocity. The number of GFP positive cells were counted automatically using the volocity software. This converts the TIFF files to grey scale images and analyses the contrasting images to the background. Each positive cell is annotated with an X and the number of positive cells are calculated per TIFF image. a. GFP positive 5TGM1-GFP cells after 6 hours of adhesion onto 14 day differentiated primary OLCs, before analysis. b. GFP positive 5TGM1-GFP cells after 6 hours of adhesion onto 14 day differentiated primary OLCs after Volocity software automated counting, where the out-put number of cells deemed positive were tagged with an X.

a

b


2.6 Knock-down of key HSC niche molecules

All equipment, materials and reagents used for all KD experiments are outlined in


2.6.1 Short hair-pin RNA gene knock-down technology

2.6.1.1 Principals of RNA interference technology

Ribonucleic acid interference (RNAi) technology was used to KD key HSC niche

molecules in the MCs. Different forms of RNAi technology are available for silencing

genes, which include silencing RNA (siRNA), shRNA and micro RNA (miRNA). Specific

sequences for siRNAs, shRNAs and miRNAs can be packaged into lentiviral particles

for effective delivery into cells. Short hair-pin RNAs are usually designed to be

between 19 and 23 nucleotide BPs long on both the sense and anti-sense strands

with a 9 BP loop and are identical to the target mRNA sequence (188).

Lentiviral particles are commonly produced in HEK293T cells, a type of renal epithelial

cell line, which are notably easy to transfect. In third generation shRNA containing

lentiviral particles, these cells are transfected with 3-4 vectors, which contain:

1. A transgene expression cassette vector, which contains packaging genes as

well as the shRNA sequence and resistance genes e.g. pLu-green

2. A packaging vector, which encodes packaging genes such as pol, gag, rev, tat

3. An envelope expression vector, required for the formation of the viral

envelope encoding for the G-protein vesicular stomatitis virus envelope gene

The viral particles are then produced by the HEK293T cells and collected in the cell

supernatants (189), as shown in Figure 2.15. The lentiviral particles are transduced

into the host cells where the shRNA is integrated into the hosts DNA. The shRNA is


then transcribed by Polymerase II or III and translocated from the nucleus to the

cytoplasm where the hairpin loop is cleaved by an enzyme known as dicer to form

siRNA. The siRNA is then incorporated into the RNA induced silencing complex (RISC),

which cleaves the double stranded (Ds) shRNA to single stranded siRNA and guides

the single strand towards the target mRNA (190). The binding of siRNAs to their

complementary mRNA results in degradation of the complementary mRNA strand by

the enzyme argonaute-2, as shown in Figure 2.15. This therefore, removes any mRNA

for the GOI and stops subsequent translation to protein (191). As the shRNA is

directly integrated into the genome, there is permanent silencing of the gene

compared to siRNA, which has a short half-life, is also easily degraded and less than

1% of the siRNA duplex is usually present 48 hours after transduction (191).

Figure 2.15: The principals of lentiviral short hair-pin RNA, to knock-down N-cad and CXCR4 in the 5TGM1-GFP cells. a. The production of lentiviral particles from HEK293T cells. HEK293T cells can be transfected with plasmids for the transgene, which contains the shRNA and resistance sequences, a packaging vector and the enveloped cassette. The particles produced are then collected for use from the cell supernatant. b. The transduction of the lentiviral shRNA particles into the MCs to KD the GOI. The lentiviral particles are transduced into the MC and the transgene is incorporated into the hosts DNA. The shRNA is transcribed and transported from the nucleus to the cytoplasm. The hairpin is then removed by the enzyme dicer to produce siRNA. The siRNA then enters the RISC complex where the double stranded (ds) RNA is cleaved to form a single strand, which binds to the complementary mRNA, resulting in degradation by argonaute-2. This removes the mRNA and prevents protein translation. Adapted from Manjunath et al (189), permission from The Journal of Advanced Drug Delivery Reviews 30.07.14.

a b


2.6.2 Puromycin killing curves

2.6.2.1 Principals of puromycin selection

Puromycin is an aminonucleoside antibiotic, derived from the bacterium

Streptomyces alboniger which is able to kill cells via disruption of protein synthesis.

However, if the cells possess a puromycin n-acetyle-transferase (PAC) gene, they are

able to produce PAC, which is able to inactivate puromycin; resulting in viable cells.

Prior to conducting GFP transduction or KD experiments, puromycin killing curves

were conducted to establish the lowest dose of puromycin required to kill 100% of

cells within four days of treatment. The optimum dose was then used to select for

cells containing the transduced PAC gene, acquired from the control or shRNA vector.

2.6.2.2 Puromycin killing curve methodology

5T33-WT or 5TGM1-GFP cells were seeded at a density of 1x105 cells into a 12 well

plate in 2 ml complete RPMI media and incubated for 24 hours. The following day,

the cells were centrifuged and re-suspended in 2 ml of fresh complete RPMI media

followed by treatment with 0 (PBS), 0.5, 1.0, 2.5, 5.0, 7.5 and 10 µg/ ml of puromycin.

After 2 and 4 days following treatment, the viability of the cells were tested by

centrifuging the cells and re-suspending them in 300 µl of PBS before staining the

dead cells with TO-PRO-3 as described in Section 2.4.1.2. The lowest dose required

for 100% cell death over a 4-day period was then established by analysing the

number of viable cells remaining at the different concentrations of puromycin at the

2 and 4 day time points.

2.6.3. Polybrene optimisation

The reagent polybrene was used to aid the binding of the viral capsid to the cells

being transduced by neutralising the charge on the sialic acid present on the cell

membrane. The optimum concentration of polybrene per the data sheet was 5


µg/ml, however, as some concentrations of polybrene can be toxic in certain cells

optimisation was required.

2.6.3.1 Polybrene optimisation methodology

The effect of polybrene upon cell viability after 24 hours was established before use

in the GFP transduction or KD experiments. Before use, polybrene was immediately

diluted to 1 mg/ml in sterile H20 and frozen in 1 ml aliquots, which were thawed as

required. For polybrene optimisation, the 5T33-WT and 5TGM1-GFP cells were

counted and seeded into a 48 well plate at a density of 5x104 in 500 µl complete RPMI

media containing 0, 2.5, 5.0, 7.5 and 10 µg/ml polybrene and incubated for 24 hours.

After 24 hours, the number of cells present in each well were counted to establish

whether the polybrene had prevented proliferation or caused cell death after 24

hours. Viability staining could not be used in this instance as polybrene treatment

resulted in false positives by allowing the DNA binding dye to enter viable cells.

2.6.4 Green fluorescent protein transduction in 5T33-WT cells and

knock-down of CXCR4 and N-cad in 5TGM1-GFP cells

2.6.4.1 Green fluorescent protein transduction and knock-down of CXCR4 and N-

cad in the 5TGM1-GFP cells methodology

To transduce GFP and to KD CXCR4 and N-cad in the MCs, a lentiviral particle

transduction system was used which contained replication incompetent viral

particles. The particles contained the shRNA sequences for CXCR4 and N-cad or

scrambled sequence for the CTRL particles or GFP sequence. In addition, they also

contained the PAC gene to enable puromycin resistance. The sequences of the

hairpin loop and subsequent siRNA are shown in Table 2.21.


Table 2.21: The hairpin loop sequences and subsequent siRNA sequences provided in the lentiviral particles, for CXCR4 and N-cad knock-down

The 5T33-WT and 5TGM1-GFP cells were seeded at a density of 2x104 cells in 100 µl

complete RPMI media containing 5 µg/ ml polybrene (optimised using the method in

Section 2.6.3.1) and the cells were transferred to a 96 well plate. To this, lentiviral

particles were added at a range of multiplicity of infections (MOIs), which is defined

as the number of transducing lentiviral particles per cell. The formula calculating the

MOI is shown below:

MOI calculations:

Total number of cells per well X desired MOI= total transducing units (TU) required

Total TU required / TU/ml in vial of lentiviral particles = volume of lentiviral particles

required.

E.g. MOI=20,000 cells x MOI 5.0= 100,000

100,000/5x106= 0.02 ml= 20 µl

Gene shRNA SiRNA Sense siRNA anti-sense

CXCR4 A

5’GATCCCCACGCCACCAACAGTCAATTCAAGAGATTGACTGTTGGTGGCGTGGT

TTTT

5’CCACGCCACCAACAGUCAAtt

5’UUGACUGUUGGUGGCGUGGtt

CXCR4 B

5’GATCCCCTCAAGATCCTTTCCAAATTCAAGAGATTTGGAAAGGATCTTGAGGTT

TTT

5’CCUCAAGAUCCUUUCCAAAtt

5’UUUGGAAAGGAUCUUGAGGtt

CXCR4 C

5’GATCCGAGACTGACCAGTCTTGTATTCAAGAGATACAAGACTGGTCAGTCTC

TTTTT

5’GAGACUGACCAGUCUUGUAtt

5’UACAAGACUGGUCAGUCUCtt

N-cad A

5’GATCCGGAACCACATGAAATTGAATTCAAGAGATTCAATTTCATGTGGTTCCTT

TTT

5’GGAACCACAUGAAAUUGAAtt

5’UUCAAUUUCAUGUGGUUCCtt

N-cad B

5’GATCCGAACCCAACTCAATTAACATTCAAGAGATGTTAATTGAGTTGGGTTCTT

TTT

5’GAACCCAACUCAAUUAACAtt

5’UGUUAAUUGAGUUGGGUUCtt

N-cad C

5’GATCCGGAAGCTGGTATCTATGAATTCAAGAGATTCATAGATACCAGCTTCCTT

TTT

5’GGAAGCUGGUAUCUAUGAAtt

5’UUCAUAGAUACCAGCUUCCtt


MOIs of 5, 15 and 25 were used and the correct volume of lentiviral particles was

added to each well. The cells were incubated for 20 hours after which they were

centrifuged and washed twice in 1 ml complete RPMI, re-suspended in 200 µl

complete RPMI and incubated. After three further days of incubation, the cells were

centrifuged and re-suspended in 500 µl complete RPMI media containing 7.5 µg/ml

puromycin for the 5T33-WT cells and 5 µg/ml puromycin for the 5TGM1-GFP cells.

The cells were then incubated until resistant colonies were visualised, approximately

5 days after selection. The 5T33-WT resistant colonies were expanded and analysed

to detect GFP expression using flow cytometry. The method for this was the same for

GFP detection in the cell lines using flow cytometry in Section 2.4.1.2. The 5TGM1-

GFP cells transduced with shRNA, were set-up for RNA extraction as described in

Section 2.3.1 1 to confirm KD of the genes. Cells were also frozen down for future

experiments as described in 2.1.2.6 or they were also isolated for WB to confirm KD

of protein, as described in Section 2.4.4.2.

The cells which had been successfully knocked-down for CXCR4 and N-cad were also

cultured in the absence of puromycin for 4 weeks to mimic the time in which they

would not be under selection when in vivo. Real-time PCR and WB were conducted

in these cells to confirm the stability of the KD.

2.6.4.2 PCR analysis

PCR analysis of the KD of CXCR4 and N-cad was performed by conducting RNA

extraction (Section 2.3.1 1), RT (Section 2.3.2.1) and real-time PCR (Section 2.3.5.2)

using the CTRL and KD cells at the various MOIs. The fold change and percentage of

KD was calculated by normalising the GOI Ct values to the HK gene β2M and the KD

was then calculated using the comparative ΔΔCt method as shown in Table 2.22.


Table 2.22: An example of calculations used to quantify the knock-down of CXCR4 and N-cad compared to CTRLs using real-time PCR

2.7. Bone parameter analysis

All equipment, materials and reagents used for bone parameter analysis are outlined

in Appendix 1, Section 1.14, Table 1.14.

2.7.1 Micro-computed tomography

2.7.1.1 Principals of micro-computed tomography

To determine any differences between the bone parameters in the mice injected

with 5TGM1-GFP cells, CTRL or KD cells, micro-computed tomography (µCT) was

used. Micro-computed tomography was introduced in the 1980s by Feldcamp et al

for the evaluation of the bone micro-architecture in mice and small rodents (192).

However, the concept of producing three-dimensional images of bone originated in

Treatment β2M N-cad

Δct ΔΔCt SD Fold change

SD %

HK GOI GOI-HK

2^-ΔCT

SD

KD/ CTRL

SD control/ control average amount

CTRL 15 15.38 23.10 7.71 0.00476492 0.00112332 1.00

0.19570940 100

CTRL 15 15.37 22.88 7.51 0.00548611

CTRL 15 15.62 22.78 7.17 0.00696819

Mean 15.46 22.92 7.46 0.00573974

SD treatment/ control amount

KD 15 16.01 24.61 8.60 0.00257122 0.00073240 0.30

0.12760174 30

KD 15 15.60 25.42 9.81 0.00111210

KD 15 15.50 24.68 9.18 0.00173001

Mean 15.70 24.90 9.20 0.00170390


1917 and was proposed by John Radon. Micro-computed tomography is a non-

invasive and non-destructive technique using X-rays. X-ray photons are generated by

accelerated electrons in the source (usually an X-ray tube) which strike the tungsten

within the tube. X-rays are emitted as a polychromatic beam, which pass through a

rotating sample within the sample holder. X-rays are then absorbed or scattered

depending on the density of the sample, dense samples such as bone which contain

calcium cause x-ray absorption whereas soft tissues result in x-ray scattering. The

images are detected and projected as two-dimensional image slices, which are

reassembled to produce three-dimensional images. The main advantage of µCT is the

spatial resolution of the machine, which can be as low as 4 µm producing highly

accurate and detailed images. The main constituents of a CT machine are shown in

Figure 2.16. A number of parameters can be analysed using µCT and the most

informative are trabecular bone volume normalised to tissue volume (BV/TV),

trabecular thickness (tb.th), trabecular number (tb.n), trabecular pattern factor

(tb.pf) and cortical thickness (ct.th). In addition, lesion number and lesion area can

be calculated.

Figure 2.16: A diagram demonstrating the main constituents of a micro-computed tomography machine. Micro-computed tomography works by producing x-rays from electrons interacting with tungsten in the x-ray tube. The x-rays emitted exist as a polychromatic beam, which passes through the collimator and specific filters to the sample in the sample holder. X-rays are then either absorbed or scattered and these differences are detected. The results can then be re-constructed using specific software to produce 3D images. Adapted from Bouxsein et al (192), permission from The Journal of Bone and Mineral Research.


2.7.1.2 Micro-computed tomography methodology

The right tibia from each mouse was fixed in 10% formalin prior to scanning. Formalin

was made using the reagents and volumes shown in Table 2.23. Bones were then

scanned using a SkyScan 1172 at 50 kilo volts (Kv), 200 microamperes (µA), using an

aluminium filter of 0.5 mm and pixel size of 4.3 µm2 and images were then

reconstructed using N-Recon software. Trabecular bone, 0.2 mm below the growth

plate, was analysed using Ct-an and Batman software to provide data for bone

parameters including BV/TV, tb.n and tb.th. All bones were scanned in-house by Holly

Evans, a trained µ-CT technician.

Table 2.23: Reagents and quantities required to make 10% formalin

Item Supplier

Sodium dihydrogen orthophosphate dihydrate 8 g Disodium hydrogen orthophosphate dihydrate 13 g

Concentrated formaldehyde (i.e. 37-41%) 200 ml dH2O 1800

2.7.1.2 Lesion counting

To count the number of lesions and lesion area, data sets for each bone were resized

from a 4.3 µm2 pixel size to 12.9 µm2. The cortical bone was hollowed out and files

were imported into Drishti I, rendering software. Using this software, a 3D model of

each bone was produced and each plane of bone (3 in total) was imaged and saved

as a separate file. These files were resized using Picassa 3 software and the files were

imported into ImageJ software. Lesions were then counted using the grey scale

automated analysis function to count the number and the area of each hole (pixels

2).


2.8. Statistics

All data throughout are presented as the mean using the standard error of the mean

(SEM) unless stated otherwise in the figure legend. All data were analysed statistically

using GraphPad Prism Version 6. Where relevant, statistical analysis was conducted

with guidance from the University of Sheffield Maths and Statistics Help group and

advice from a senior clinical scientist with expertise in statistics. Throughout all

experiments, non-parametric tests were used, as Gaussian distribution could not be

assessed, as N numbers were not above the required number of eight. Mann-

Whitney U tests were used to compare values between two groups of unpaired data

and a Wilcoxon matched pairs signed rank test was used to compare two groups of

paired data. In addition, a Kruskal-Wallis test followed by Dunn’s post test was used

to compare more than two groups of unpaired data and a Friedman test followed by

Dunn’s post test was used to compare more than two groups of paired data.

Chapter III: The in vitro expression of HSC niche molecules and their complimentary ligands by myeloma and osteoblastic cells Page 98

Chapter III: The in vitro expression of HSC niche

molecules and their complementary ligands by

myeloma and osteoblastic cells


3.1 Introduction

HSCs reside in BM microenvironmental niches in which their proximity to certain BM

constituents influences their behaviour. Their proximity to osteoblasts is thought to

inhibit HSC proliferation and promote entry into the G0 phase of cell cycle, whereas,

their proximity to the vasculature stimulates proliferation and differentiation into a

variety of cell lineages (82, 83, 85). Molecules which have been implicated in the

physical interaction between HSCs and osteoblastic cells, are the receptors CXCR4,

Notch-1, Tie-2 and the adherin molecule N-cad, expressed by the HSCs, and their

complementary ligands, CXCL12, Jag-1, Ang-1 and N-cad, expressed or secreted by

the osteoblastic cells. The BM and its cellular constituents also play an important role

in MC homing, growth and survival. Previous studies have largely focussed upon the

influence of the BM microenvironment in the later stages of disease and current

research is particularly limited with regards to early myeloma development as well

as MC survival following chemotherapy. It has been suggested that MCs may exist

within a BM microenvironmental niche, distinct but similar to that occupied by HSCs,

or, MCs may in fact “hijack” the HSC niche. As with HSCs, occupancy of the niche may

result in MC quiescence if located in close proximity to osteoblasts. Currently, little

research has been conducted to test this hypothesis. However, previous papers have

identified the expression of some of the HSC niche molecules by MCs in vitro.

There is substantial evidence in support of the CXCR4/ CXCL12 axis as part of the

mechanism for MC homing and mobilisation to the BM. Previous data have described

the in vitro expression of CXCR4 by various patient cells and human myeloma cell

lines including MM.1S, U266, OPM-2, Kas/61, RPMI-8226, KMS-12BM, KMS-12PE and

KHM-1B cells (193-195) and murine cell lines such as the 5T2MM and 5T33MMvt

(170). In addition, its ligand CXCL12 is widely known to be expressed by human

osteoblast cell lines including MBA-15, MG63 (111) and MC3T3-E1 (196) as well as

primary human osteoblasts (111) in vitro.


The interactions of Notch-1 and Jag-1 have also been investigated in MM. Notch-1

was expressed by several human MC lines in vitro including; NCI H929, RPMI-8226,

HS-Sultan, ARH-77, IM-9, MC/CAR and MM.1S (197) as well as primary human MCs

(198). The ligand Jag-1 was also expressed by human primary BM stromal cells (199)

and MC3T3-E1 osteoblastic cells (200) in vitro.

Currently there is limited data available in support of the expression of Tie-2 by MCs

compared to that published for CXCR4 and Notch-1. However, Tie-2 expression was

determined in the human cell line XG-1 and in 4 out of 23 MM patient plasma cell

samples by RT-PCR. Ang-1 expression was identified in numerous myeloma cell lines

in vitro, including RPMI-8226, U266, OPM-2, XG-1, and XG-6 as well as 11 out of 22

patient samples (201). Ang-1 was also expressed by murine primary osteoblastic

cells, ST-2 cells and C2C12 myoblastic cells (202) in vitro.

A role for N-cad in metastasis has been explored in several cancers including prostate

and breast cancer. Prostate cell lines PC-3 and LAPC4 and LAPC9 cells (203) and breast

carcinoma cell lines BT549, Hs578t, SUM159PT, MDA-MB-435 and MDA-MB-436

expressed N-cad (204). N-cad expression has also been determined in human patient

MM cells (205) and in human MC lines; LME-1, UM-1, OPM-1, NCI-H929 (206), U266

and RPMI (207) in vitro. N-cad was also expressed by C2C12 myoblasts and the MSC

progenitor cell line C3H10T1/2 (which undergo osteoblast differentiation once

treated with BMP-2 and 7), KS483 osteoblastic cells (206) as well as murine MC3T3-

E1 cells (208) in vitro.

Taken together the studies above provide evidence for a myeloma and osteoblastic

cell axis in which the MCs express the HSC niche molecules CXCR4, Notch-1, Tie-2 and

N-cad and the osteoblastic cells express the ligands CXCL12, Jag-1, Ang-1 and N-cad.


Currently there is no data, which systematically determines the expression of these

molecules in the 5TGM1-GFP cell line and limited data for the 5T33MMvt model. In

my studies, I have determined the in vitro expression profile of CXCR4, Notch-1, Tie-

2 and N-cad by the 5T33-WT, 5T33-GFP and the 5TGM1-GFP cells and the expression

of the complementary ligands CXCL12, Jag-1, Ang-1 and N-cad by murine osteoblastic

cells, cultured in vitro.

3.2 Hypothesis, aims and objectives

3.2.1. Hypothesis and aims

The aim of these studies was to test the hypothesis that “Myeloma cells express the

same repertoire of molecules as HSCs, and their complementary ligands are

expressed by osteoblastic cells in vitro”.

3.2.2 Objectives

The hypothesis will be tested by the following objectives:

1. To determine the expression of CXCR4, Notch-1, Tie-2 and N-cad by the 5T33-

WT, 5T33-GFP and 5TGM1-GFP cell lines using endpoint and real-time PCR

and to evaluate the presence of each protein by flow cytometry,

immunofluorescence (IF) and Western blotting (WB).

2. To determine the expression of CXCL12, Jag-1, Ang-1 and N-cad by the

MC3T3-E1 cells and murine primary osteoblast lineage cells (OLCs) using

endpoint and real-time PCR and to evaluate the presence of each protein by

flow cytometry.


3.3 Methods

3.3.1 Gene expression analysis

3.3.1.1 Endpoint PCR

Murine BM was isolated from C57BL/6 mice (N=3) as described in Chapter II, Section

2.2.1. The 5T33-WT, 5T33-GFP, 5TGM1-GFP, MC3T3-E1 cells and primary OLCs were

prepared 24 hours prior to RNA extraction to ensure experimental consistency. One

million MCs were seeded in triplicate into a 6 well plate with 3 ml complete RPMI

media. The MC3T3-E1 cells and primary OLCs were seeded in triplicate at a density

of 1x106 into a T75 flask in 12 ml of complete media, appropriate for the cell type.

Each cell type was incubated for 24 hours before RNA extraction, followed by cDNA

synthesis and PCR as described in sections 2.3.1, 2.3.2, 2.3.3 and 2.3.4. The process

from seeding the cells to PCR was repeated three times to produce three biological

replicates. The Genomics Facility at The University of Sheffield then sequenced all

PCR products and the homology was analysed by comparing the sequencing results

with the expected template and GenBank archives.

3.3.1.2 Real-time PCR

The cDNA from each cell type, generated as described above was also used for real-

time PCR analysis as described in section 2.3.5. Expression was analysed using an

Applied Biosystems 7900HT real-time PCR machine to produce a Ct value, which was

normalised to the HK gene β2M to produce the ΔCt value. As with endpoint PCR, the

process from RNA extraction to PCR was repeated three times to produce three

biological replicates (N=3).

3.3.2 Protein quantification analysis

3.3.2.1 Flow cytometry

The murine BM, splenocytes and blood were isolated and processed for flow

cytometry as described in section 2.2.1 and the MCs and murine osteoblastic cells

were set-up as described above for RNA extraction (N=3-6). Cells were then

processed for flow cytometry as described in section 2.4.1. The percentage of

positive viable cells (using PI or TO-PRO-3 staining) was measured using a FACS

Calibur and analysed using Cell Quest software by gating against the negative isotype

control, which demonstrated less than 1% positive staining. Each cell type was set-

up in triplicate and repeated three times (N=3) to produce a mean percentage of cells

positive for each molecule.

3.3.2.2 Immunofluorescence

Immunofluorescence was used to quantify and visualise the protein for the HSC niche

molecules, ligands and CD138 by the 5T33-GFP and 5TGM1-GFP cells as described in

section 2.4.2. Positive cells were visualised using a Leica AF6000LX microscope, a 40X

magnification and LAS LF software. Analysis was conducted using Volocity software

to calculate the percentage of cells, which were positive for each molecule. Each

experiment was set-up in triplicate (using three slides) and repeated three times

(N=3) to calculate a mean percentage of cells which were positive for each molecule.

3.3.3. Statistics

Where relevant non-parametric statistical tests were used. A Kruskal-Wallis test

followed by a Dunn’s post test (P<0.05) was used to determine any differences in

gene expression between the MCs and the presence of protein for the HSC niche

molecules and ligands between the murine constituents and MCs. Where relevant,

error bars were displayed on all graphs using the SEM unless stated otherwise in the

legend.


3.4 Results

3.4.1 Qualitative assessment of gene expression for the HSC niche molecules,

ligands and CD138 by normal murine constituents and murine multiple myeloma

cells, determined using endpoint PCR

Endpoint PCR was used to determine the qualitative gene expression of the HSC

niche molecules, their ligands and the plasma cell marker CD138 by murine BM (to

act as a positive control for the primers), 5T33-WT, 5T33-GFP and 5TGM1-GFP cells.

As demonstrated in Figure 3.1, endpoint PCR of the murine BM generated products

of expected size for the HK gene GAPDH (354 BP) and the HSC niche molecules; CXCR4

(711 BP), Notch-1 (348 BP), Tie-2 (648 BP) and N-cad (560 BP) and ligands; CXCL12

(205 BP), Jag-1 (370 BP), Ang-1 (401 BP) and CD138 (559 BP). These results therefore,

confirmed that endpoint PCR primers were working effectively.

Endpoint PCR was also used to determine the qualitative gene expression of the HSC

niche molecules, ligands and CD138 by the MCs. Endpoint PCR of the 5T33-WT, 5T33-

GFP and 5TGM1-GFP cells generated products of expected size for the HK gene

GAPDH and the HSC niche molecules CXCR4 and Notch-1. In addition, the 5T33-WT

and 5TGM1-GFP samples also generated a band for N-cad but the 5T33-GFP cells did

not (Figure 3.1). A band for CD138 was visualised in each MC type. Endpoint PCR

analysis of MC lines did not generate products for the ligands CXCL12, Jag-1, Ang-1

or the receptor Tie-2. Therefore, the MCs expressed some HSC niche molecules and

no ligands by endpoint PCR.


Figure 3.1: Murine bone marrow and myeloma cells expressed HSC niche molecules, ligands and CD138 by endpoint PCR. Endpoint PCR reactions were run using approximately 40 ng of RT+ cDNA samples. An equal volume of RT- and H2O control samples were used to assess genomic and reagent contamination. The PCR reactions were run for 25 cycles for GAPDH and 35 cycles for the HSC niche molecules, ligands and CD138. The PCR products were separated by electrophoresis in a 1.6% agarose gel and visualised using a Bio-rad Gel Doc XR+ Imager system and Quantity One 4.6.8 software. Each row depicts an individual gel electrophoresis for GAPDH and each HSC molecule and ligand. The base pair (BP) number on the right corresponds to the size of bands visualised and band products were confirmed by in-house sequencing. Results are representative of three independent experiments.

3.4.2 Quantitative assessment of gene expression for the HSC niche molecules,

ligands and CD138 by normal murine constituents and murine multiple myeloma

cells, determined using real-time PCR

Real-time PCR was used to determine the quantitative expression of the HSC niche

molecules, their ligands and CD138 by the murine BM, 5T33-WT, 5T33-GFP and

5TGM1-GFP cells. Table 3.1 demonstrates a summary of the absolute gene

expression values (Ct values) generated using real-time PCR. The BM expressed the


HK genes β2M and HPRT, HSC niche molecules CXCR4, Notch-1, Tie-2 and N-cad and

ligands CXCL12, Jag-1 and Ang-1. In addition, they also expressed CD138. This

confirmed that all probes were effective.

Table 3.1 also shows the absolute gene expression of the HK, HSC niche molecules

and ligands by the 5T33-WT, 5T33-GFP and 5TGM1-GFP cells and in addition, Figure

3.2 shows the 1/ΔCt values relative to the HK gene β2M. The 5T33-WT, 5T33-GFP and

5TGM1-GFP samples highly expressed the HK genes β2M and HPRT, the HSC niche

molecules CXCR4 and Notch-1 and plasma cell marker CD138 but demonstrated low

expression of the ligand CXCL12. The 5T33-WT expressed N-cad, as did the 5TGM1-

GFP, though at a higher level, whereas, N-cad was undetermined (UD) in the 5T33-

GFP samples. The 5TGM1-GFP samples also had marginal expression of Tie-2 and

Ang-1, which were absent in both the 5T33-WT and 5T33-GFP samples. Each cell type

did not express the ligand Jag-1. When the data was analysed statistically to compare

the expression between the MCS, there were no significant differences. However, it

was clear from the data that N-cad expression was higher in the 5TGM1-GFP cells

compared to both the 5T33-WT and 5T33-GFP cells but because some of the values

were UD this was unable to be analysed statistically.

In summary, as with the endpoint PCR analysis, the MCs expressed only some of the

HSC niche molecules and ligands using real-time PCR and the pattern of expression

was broadly similar to the endpoint PCR analysis where absent expression

corresponded with high or UD Ct values.

Table 3.1: A summary of the absolute gene expression of the HSC niche molecules, ligands and CD138 by murine bone marrow and myeloma cells using real-time PCR

Gene BM 5T33-WT 5T33-GFP 5TGM1-GFP

β2M 17 ± 0.16 15 ± 0.11 15 ± 0.06 12 ± 0.22 HPRT 22 ± 0.17 20 ± 0.24 20 ± 0.04 17 ± 0.17 CXCR4 21 ± 0.08 21 ± 0.05 21 ± 0.25 19 ± 0.08

Notch-1 23 ± 0.23 24 ± 0.09 25 ± 1.38 22 ± 0.34 Tie-2 29 ± 0.07 UD UD 34 ± 0.56 N-cad 27 ± 0.44 33 ± 0.33 UD 20 ± 0.06

CXCL12 21 ± 0.55 31 ± 0.99 33 ± 1.38 30 ± 0.51 Jag-1 28 ± 0.51 UD UD UD Ang-1 27 ± 0.17 UD UD 34 ± 0.75 CD138 28 ± 0.81 21 ± 0.06 21 ± 0.10 18 ± 0.20

Mean Ct values ± SEM UD= Undetermined, Ct value > 35.

Figure 3.2: The murine bone marrow and myeloma cells expressed the HSC niche molecules and complementary ligands at varying levels using real-time PCR analysis. Gene expression was assessed by normalising the Ct values produced in real-time PCR for CXCR4 & CXCL12 (a), Notch-1 & Jag-1 (b), Tie-2 & Ang-1 (c) and N-cad & CD138 (d) to the HK gene β2M to produce the 1/ΔCt (y-axis) value. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s post test (P<0.05) to compare expression between the MCs. Results were determined from three independent experiments (N=3) using the SEM.

a b

c d

3.4.3 Protein quantification of HSC niche molecules, ligands and CD138 by normal

murine constituents and murine multiple myeloma cells using flow cytometry

Flow cytometry was used to determine the number of murine BM, splenocytes, blood

cells and MCs, which were positive for the HSC niche molecules, ligands and CD138.

Figure 3.3a demonstrates an example of the flow cytometry plots produced for the

staining of CD138 and the viability dye PI in samples of control BM, splenocytes and

peripheral blood. The mean percentage of cells which were positive for each HSC

niche molecule and ligand (excluding Ang-1 and N-cad) are shown in Figure 3.3.b and

summarised in Table 3.2. The BM, splenocytes and blood were positive for the HSC

niche molecules; CXCR4, Notch-1 and Tie-2 and the ligands; CXCL12 and Jag-1. CD138

positive staining was also detected in the BM, splenocytes and blood. When the

results were compared statistically, a significantly greater number of BM cells were

positive for CXCR4 compared to the splenocytes (P<0.05) and there were no other

statistically significant differences between the remaining HSC niche molecules and

ligands when comparing the different cell types.

Ang-1 was not analysed by flow cytometry due a lack of directly conjugated antibody

and unsuccessful use of an indirect antibody method using the murine constituents

(data not shown). In addition, when the work was originally conducted there were

no directly conjugated N-cad antibodies available and indirect methods failed to stain

the murine positive controls (data not shown). However, more recently a directly

conjugated N-cad antibody became available and this was tested using the 5TGM1-

GFP cells and the 5T33-GFP cells as a negative control as shown later in Figure 3.5.

Figure 3.3: The control murine BM, splenocytes and blood samples were positive for the HSC niche molecules, complementary ligands and CD138 by flow cytometry. Murine BM, splenocytes and blood cells were incubated with directly conjugated primary antibodies for the HSC niche molecules; CXCR4, Notch-1 and Tie-2, ligands; CXCL12 and Jag-1 and CD138 or dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. a. An example of a representative scatter plot for CD138 using the isotype or primary antibody for the murine BM (i-ii), splenocytes (iii-iv) and blood (v-vi). Positive staining for CD138 and negative PI staining is shown in the bottom right hand side of the quadrant. b. Quantitative analysis for the percentage of viable cells which were positive for CXCR4/CXCL12, (i), Notch-1/Jag-1 (ii) and Tie-2 & CD138 (iii) by the BM, splenocytes and blood cells. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s post test (P<0.05). Results were determined from three to six animals (N=3-6).

i iii v a

bi

ii iv vi

ii

iii

Table 3.2: A summary of the quantitative analysis for the percentage of control murine constituents, which were positive for the HSC niche molecules, ligands and CD138 using flow cytometry

Protein BM Splenocytes Blood

CXCR4 65.75 ± 5.08 37.24 ± 4.70 49.06 ± 3.04 Notch-1 2.79 ± 0.81 4.87 ± 0.69 3.00 ± 2.98

Tie-2 11.95 ± 2.37 11.09 ± 4.34 1.89 ± 0.16 CXCL12 11.29 ± 2.39 5.44 ± 1.12 7.19 ± 3.09

Jag-1 13.18 ± 13.18 1.48 ± 1.48 4.57 ± 0.20 CD138 10.01 ± 4.03 6.68 ± 3.36 7.87 ± 3.06

Mean percentage of positive cells ± SEM

Figure 3.4.a. shows an example of the flow cytometry plots produced for the staining

of CD138 and the viability dye PI in the 5T33-WT, 5T33-GFP and 5TGM1-GFP cells.

The mean percentage of cells which were positive for each HSC niche molecule and

ligand are shown in Figure 3.4.b. Table 3.3 also summaries the percentages and SEM

values for the mean number of 5T33-WT and GFP and 5TGM1-GFP cells which were

positive for the HSC niche molecules, ligands and CD138. Each MC was positive for

the HSC niche molecules CXCR4, CXCL12, Notch-1, Tie-2 and CD138. In addition, the

5TGM1-GFP cells were positive for the ligand Jag-1. When the proportion of positive

cells were compared between the MCs statistically, a greater number of 5T33-WT

cells were positive for CXCR4 (P<0.05) and Notch-1 (P<0.05) compared to the 5T33-

GFP cells. In addition a greater number of 5TGM1-GFP cells were positive for CXCL12

(P<0.05), Notch-1 (P<0.05) and Jag-1 (P<0.05) compared to the 5T33-GFP cells.

Figure 3.4: The MM cells were positive for the HSC niche molecules, some of the complementary ligands and CD138 at varying degrees using flow cytometry. 5T33-WT, 5T33-GFP and 5TGM1-GFP cells were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4, Notch-1 and Tie-2, ligands CXCL12 and Jag-1 and, CD138 or dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. a. An example of a representative scatter plot for CD138 using the isotype control and primary antibody by the 5T33-WT (i-ii), 5T33-GFP (iii-iv) and 5TGM1-GFP cells (v-vi). Positive staining for CD138 and negative PI staining is shown in the bottom right hand side of the quadrant. b. Quantitative analysis for the percentage of viable cells which were positive for CXCR4/ CXCL12 (i), Notch-1/Jag-1 (ii) and Tie-2 & CD138 (iii) by the 5T33-WT, 5T33-GFP and 5TGM1-GFP cells. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s post test (P<0.05) Results were determined from three independent experiments (N=3).

i iii v a

bi

ii iv vi

iii

ii


Table 3.3: A summary of the quantitative analysis for the percentage of myeloma cells, which were positive for the HSC niche molecules, ligands and CD138 using flow cytometry

Protein 5T33-WT 5T33-GFP 5TGM1-GFP

CXCR4 97.19 ± 0.06 90.59 ± 0.32 94.79 ± 1.41 Notch-1 96.45 ± 0.10 64.43 ± 1.68 97.91 ± 0.26

Tie-2 4.15 ± 0.67 0.34 ± 0.07 4.97 ± .0.78 CXCL12 32.19 ± 4.15 3.76 ± 1.40 64.20 ± 3.00

Jag-1 0.10 ± 0.03 0.03 ± 0.01 6.99 ± 1.18 CD138 89.56 ± 0.93 89.59 ± 0.38 98.65 ± 0.20


As stated previously, in more recent experiments a directly conjugated N-cad

antibody was used to stain the 5T33-GFP (used as a negative control due to the

endpoint and real-time PCR data) and 5TGM1-GFP cells for N-cad. A dose matched

isotype control was also used however; the isotype control did not have the same

conjugate as the primary antibody (Dylight 650) as this was unavailable. Therefore,

the results between the isotype and primary antibody are not completely

comparable. Figure 3.5 demonstrates the percentages of 5T33-GFP and 5TGM1-GFP

cells, which were positively stained for N-cad using increasing amounts of primary

antibody. As demonstrated in the graph, there was an increase in the number of

positive 5T33-GFP and 5TGM1-GFP cells as the amount of primary antibody increased

and using 1.0 µg of antibody, nearly all of the 5T33-GFP and the 5TGM1-GFP cells

were positively stained (98% and 96% respectively). In addition, there were no

differences between the number of 5T33-GFP or 5TGM1-GFP cells, which were

positive for N-cad, when the percentages were examined. As these data were

contradictory to the endpoint and real-time PCR data, where 5T33-GFP cells did not

show any expression of N-cad, the reliability of the antibody is questionable.


Figure 3.5: An increase in the concentration of N-cad antibody correlated with an increase in positive staining in both the 5T33-GFP and 5TGM1-GFP cells using flow cytometry. 5T33-GFP and 5TGM1-GFP cells were incubated with increasing doses of primary conjugated antibodies for N-cad or dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. Quantitative analysis for the percentage of positive viable cells for N-cad by the 5T33-GFP and 5TGM1-GFP was conducted. No statistical analysis was conducted. Results were determined from three independent experiments (N=3).

3.4.4 N-cad protein detection by 5T33-GFP and 5TGM1-GFP cells using Western

blotting

As the flow cytometry data for N-cad did not seem reliable due to the linear increase

in positive staining as the amount of antibody increased using the 5T33-GFP and

5TGM1-GFP cells, WB was used as an alternative. This was conducted using a

different unconjugated primary antibody, described in Section 2.4.4.

Figure 3.6 demonstrates the WB produced to detect N-cad in the 5T33-GFP and

5TGM1-GFP lysates. Two different ECL solutions were used, one was purchased from

Thermo Scientific, as shown in Figure 3.6a and b respectively and the other was

purchased from Santa Cruz Biotechnology as shown in Figure 3.6c and d. The ECL

purchased from Thermo Scientific, despite a significantly shorter exposure time than

the Santa Cruz Biotechnology ECL (30 seconds), over-exposed the protein bands,


particularly in the protein ladder, which made the ladder difficult to interpret.

Whereas, the Santa Cruz Biotechnology ECL which needed a longer exposure time (5

min), produced clearer WBs with a distinguishable protein ladder and clear bands for

β-actin and N-cad. Therefore, this ECL was the most desirable to use in future

experiments.

Figure 3.6: N-cad protein was present in the 5TGM1-GFP cells but not in the 5T33-GFP cells using Western blotting. 5T33-GFP or 5TGM1-GFP cell lysates (20 µg) or biotinylated protein ladder (1:10) were run in a 7% polyacrylamide gel and the separated proteins were transferred on to a nitrocellulose membrane. The membranes were blotted with primary antibodies for N-cad and β-actin followed by HRP-conjugated secondary antibodies and an anti-biotin-HRP antibody. The blots were developed using ECL supplied by Thermo Scientific (exposure time of 30 seconds) and Santa Cruz Biotechnology (exposure time of 5 min). The experiment was repeated three times using three biological lysate replicates and representative images were produced. As shown above, lane 1 demonstrates the protein ladder and lane 2 shows the cell lysates. Proteins are labelled and the size of the protein is labelled on the left (kDa) a. 5T33-GFP lysates and Thermo Scientific ECL. b. 5TGM1-GFP lysates and Thermo Scientific ECL. c. 5T33-GFP lysates and Santa Cruz Biotechnology ECL. d. 5TGM1-GFP lysates and Santa Cruz Biotechnology ECL.

Figure 3.6 also demonstrates differences in the presence of N-cad between the 5T33-

GFP (Figure 3.6a and c) and 5TGM1-GFP cells (Figure 3.6b and d). Using both the

Thermo Scientific ECL and Santa Cruz ECL, clear β-actin bands were produced of the

correct size for both the 5T33-GFP and 5TGM1-GFP lysates. However, there was no

band for N-cad using the 5T33-GFP lysates whereas, clear bands of the correct size

a b

c d

Protein ladder

Lane 1 Lane 2

Lane 1 Lane 2 Lane 1 Lane 2

Lane 1 Lane 2

5T33-GFP

5T33-GFP

5TGM1-GFP

5TGM1-GFP


were detected using the 5TGM1-GFP lysates. Therefore, this implies that N-cad

protein is not present in the 5T33-GFP cells but is present in the 5TGM1-GFP cells.

3.4.5 Protein visualisation of HSC niche molecules, ligands and CD138 by 5T33-GFP

and 5TGM1-GFP cells using immunofluorescence

Immunofluorescence was used to confirm the presence of protein for the HSC niche

molecules, ligands and CD138 by the 5T33-GFP and 5TGM1-GFP cells as well as to

visualise the protein within the cells. Figures 3.7 and 3.8 demonstrate the images

taken of 5T33-GFP and 5TGM1-GFP cells respectively, stained for CXCR4, Notch-1,

CXCL12 and CD138. Images of Tie-2 and Jag-1 were not shown, as primary antibody

staining was not apparent in either cell type compared to the isotype control. From

the images, it was clear that both cell types were positive for CXCR4, CXCL12, Notch-

1 and CD138 and the positive staining appeared to locate to specific regions of the

cells depending on the molecule. CXCR4 and CD138 staining was visualised on the

membrane of the cells whereas Notch-1 seemed to be stained within membranous

and cytoplasmic regions and CXCL12 was present inside the cell cytoplasm as discrete

foci.


Figure 3.7: 5T33-GFP cells were positive for HSC niche molecules CXCR4, Notch-1, the ligand CXCL12 and the plasma cell marker CD138 by immunofluorescence. 5T33-GFP cytospins were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4 and Notch-1, the ligand CXCL12 and CD138 (right-hand side) and dose matched isotype control antibodies (left-hand side). This was followed by visualisation using a Leica AF6000LX microscope at a 40X magnification using LAS LF software. Nuclei are blue (DAPI) and positive antibody staining is red (APC). Images are representative of three independent experiments. Scales bars represent 50 µM.

CXCR4

Notch-1

CXCL12

CD138

Isotype antibody Primary antibody Primary antibody


Figure 3.8: 5TGM1-GFP cells were positive for HSC niche molecules CXCR4, Notch-1, the ligand CXCL12 and the plasma cell marker CD138 by immunofluorescence. 5TGM1-GFP cytospins were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4 and Notch-1, the ligand CXCL12 and CD138 (right-hand side) and dose matched isotype control antibodies (left-hand side). This was followed by visualisation using a Leica AF6000LX microscope at a 40X magnification using LAS LF software. Nuclei are blue (DAPI) and positive antibody staining is red (APC). Images are representative of three independent experiments. Scales bars represent 50 µM.

Using the images generated from IF, the percentage of 5T33-GFP and 5TGM1-GFP

cells, which were positive for each molecule, was calculated, shown graphically in

Figure 3.9 and summarised in Table 3.4. From the data, there were no differences

between the number of cells 5T33-GFP and 5TGM1-GFP cells, which were positive

CXCR4

Notch-1

CXCL12

CD138

Isotype antibody Primary antibody Primary antibody


for CXCR4, Notch-1 or CD138, but a greater number of 5TGM1-GFP cells, were

positive for CXCL12 compared to the 5T33-GFP cells.

Figure 3.9: The 5T33-GFP and 5TGM1-GFP cells were positive for the HSC niche molecules CXCR4 and Notch-1, ligand CXCL12 and CD138 using quantitative analysis of immunofluorescence staining. 5T33-GFP and 5TGM1-GFP cytospins were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4 (a) and Notch-1 (c), the ligand CXCL12 (b) and CD138 (d) or dose matched isotype control antibodies. This was followed by visualisation using a Leica AF6000LX microscope at a 40X magnification and LAS LF software. Volocity spot counting software was then used to calculate the mean percentage of cells, which were positive for each molecule. Results were determined from three independent experiments (N=3).

a b

c d


Table 3.4: A summary of the quantitative analysis for the percentage of 5T33-GFP and 5TGM1-GFP cells, which were positive for the HSC niche molecules; CXCR4 and Notch-1, ligand CXCL12, and CD138 using immunofluorescence

Protein 5T33-GFP 5TGM1-GFP

CXCR4 95.70 ± 1.21 96.33 ± 0.34 Notch-1 98.70 ± 0.89 98.67 ± 0.34 CXCL12 20.30 ± 3.78 71.00 ± 5.55 CD138 97.00 ± 0.58 99.90 ± 1.01


3.4.6 Qualitative assessment of gene expression for the HSC niche molecules and

ligands by MC3T3-E1 and primary osteoblast lineage cells using endpoint PCR

Endpoint PCR was used to determine the qualitative gene expression of the HSC

niche molecules and their ligands by the murine MC3T3-E1 cells and murine primary

osteoblast lineage cells (OLCs). As demonstrated in Figure 3.10, both the MC3T3-E1

cells and primary OLC samples produced bands for ligands; CXCL12, Jag-1 and Ang-1,

and HSC niche molecules; Notch-1 and N-cad. In addition, the primary OLCs produced

a band for Tie-2. Therefore, the osteoblastic cells were positive for all ligands and

some of the HSC niche molecules using endpoint PCR.


Figure 3.10: The murine osteoblastic cells expressed the complementary ligands and some of the HSC niche molecules using endpoint PCR. The endpoint PCR reactions were run using approximately 40 ng of RT+ cDNA samples. An equal volume of RT- and H2O control samples were used to assess genomic and reagent contamination. The PCR reactions were run at 25 cycles for GAPDH and 35 cycles for the HSC niche molecules and ligands. The PCR product was separated by electrophoresis in a 1.6% agarose gel and visualised using a Bio-rad Gel Doc XR+ Imager system and Quantity One 4.6.8 software. Each row depicts an individual gel electrophoresis for GAPDH and each HSC molecule and ligand. The base pair (BP) number on the right corresponds to the size of bands visualised and band product was confirmed by in-house sequencing. Results are representative of three independent experiments.

3.4.7 Quantitative assessment of gene expression for the HSC niche molecules and

ligands by MC3T3-E1 and primary osteoblast lineage cells using real-time PCR

Real-time PCR was used to determine the quantitative expression of the HSC niche

molecules and their ligands by MC3T3-E1 cells and OCLs. Table 3.5 demonstrates a

summary of the absolute gene expression values (Ct values) and Figure 3.11 shows

the 1/ΔCt values generated using real-time PCR.


Table 3.5: A summary of the absolute gene expression for the HSC niche molecules and ligands by MC3T3-E1 cells and primary osteoblast lineage cells using real-time PCR

Gene MC3T3-E1 Primary OLCs

β2M 19 ± 0.94 17 ± 0.12 HPRT 25 ± 0.90 23 ± 0.08 CXCR4 UD UD

Notch-1 28 ± 1.26 25 ± 0.34 Tie-2 UD 29 ± 0.37 N-cad 24 ± 1.28 23 ± 0.16

CXCL12 22 ± 1.18 17 ± 0.21 Jag-1 25 ± 0.98 22 ± 0.13 Ang-1 28 ± 1.31 25 ± 0.16

Mean Ct values ± SEM UD= Undetermined, Ct value > 35.

Both cell types expressed the HK genes β2M and HPRT as well as the ligands CXCL12,

Jag-1 and Ang-1. In addition, they both expressed the HSC niche molecules Notch-1

and N-cad. The primary OLCs, also expressed Tie-2, whereas this was absent in the

MC3T3-E1 cells. Both cell types did not express CXCR4. It is clear from these data that

the primary OLCs expressed CXCL12 at a greater level than the MC3T3-E1 cells and

also had slightly higher expression of Notch-1, Jag-1, Tie-2 and Ang-1, whereas, N-

cad was more highly expressed in the MC3T3-E1 cells, however, statistical analysis

was unable to be conducted.


Figure 3.11: The murine osteoblastic cells expressed the complementary ligands and some of the HSC niche molecules using real-time PCR. Gene expression was assessed by normalising the Ct values produced in real-time PCR for CXCR4 & CXCL12 (a), Notch-1 & Jag-1 (b), Tie-2 & Ang-1 (c) and N-cad (d) to the HK gene β2M to produce the 1/ΔCt value (y-axis). Results were determined from three independent experiments (N=3) using the SEM.

3.4.8 Protein quantification of the HSC niche molecules and ligands by MC3T3-E1

and primary osteoblast lineage cells using flow cytometry

Flow cytometry was used to determine the number of MC3T3-E1 cells and primary

OLCs, which were positive for the HSC niche molecules and their ligands. Figure 3.12a

demonstrates an example of the flow cytometry plots produced for the staining of

Jag-1 and the viability dye TO-PRO-3 using the MC3T3-E1 cells and primary OLCs. The

mean percentage of cells, which were positive for each HSC niche molecule and

ligand, are shown in Figure 3.12b. A summary of the percentages of cells which were

positive for the HSC niche molecules and ligands is shown in Table 3.6. The MC3T3-

E1 cells and primary OLCs were positive for the ligands CXCL12 and Jag-1 and the HSC

molecule Notch-1 and neither cell type was positive for Tie-2. A slightly higher

number of primary OLCs were positive for CXCL12 compared to the MC3T3-E1 cells,

whereas, a greater number of MC3T3-E1 cells were positive for Notch-1 and Jag-1

a b

c d


compared to the primary OLCs, where the difference was particularly obvious for Jag-

1, however, statistical analysis was unable to be conducted.

Figure 3.12: The murine osteoblastic cells were positive for the complementary ligands CXCL12 and Jag-1 and HSC niche molecule Notch-1 using flow cytometry. The MC3T3-E1 cells and primary OLCs were incubated with directly conjugated primary antibodies for the ligands CXCL12 and Jag-1 and HSC niche molecules Notch-1 and Tie-2 or dose matched isotype control antibodies. This was followed by flow cytometry using a BD Calibur and Cell Quest software. a. An example of a representative scatter plot for the isotype and primary antibody for Jag-1 by the MC3T3-E1 cells (i-ii) and primary OLCs (iii-iv). Positive staining of Jag-1 and negative TO-PRO-3 is shown in the bottom right-hand side of the quadrant. b. Quantitative analysis for the percentage of viable cells, which were positive for CXCR4/CXCL12 (i) and Notch-1/Jag-1 (ii) by the MC3T3-E1 and primary OLCs. Results were determined from three independent experiments (N=3)

a

ii

bi

ii

i

iv

iii


Table 3.6: A summary of the quantitative analysis for the percentage of MC3T3-E1 cells and primary osteoblast lineage cells, which were positive for the HSC niche molecules and ligands using flow cytometry.

Protein MC3T3-E1 Primary OLCs

CXCL12 63.10 ± 3.10 75.53 ± 1.37 Jag-1 54.70.± 0.95 12.63 ± 1.01

Notch-1 82.48 ± 0.75 61.03 ± 6.64 Tie-2 0.00 ± 0.00 0.00 ± 0.00

Mean percentage of positive cells ± SEM.

3.5 Discussion

The aim of this chapter was to determine whether the HSC niche molecules CXCR4,

Notch-1, Tie-2 and N-cad were expressed by the 5T33MMvt and 5TGM1-GFP cell

lines, and to determine the expression of their complementary ligands CXCL12, Jag-

1, Ang-1 and N-cad by the MC3T3-E1 cells and murine primary OLCs in vitro.

Firstly, murine BM cDNA was used in PCR experiments, and murine BM, splenocytes

and blood cells were used for flow cytometry as positive controls to ensure that the

PCR and flow cytometry reagents and protocols were working effectively. The

expression of each HSC niche molecule and ligand by the murine BM was

demonstrated in my studies, using each methodology. However, when comparing

the BM PCR and flow cytometry data, the number of positive cells did not correlate

with the gene expression. For example there was high expression of CXCL12 and

Notch-1 using PCR, whereas the number of cells positive for CXCL12 and Notch-1 was

below 12% and 3% respectively using flow cytometry. This difference may potentially

be due to the preparation of the samples i.e. the RBCs were lysed in the flow

cytometry samples whereas the entire BM population was used for RNA extraction,

or it may also be due to inefficiency of the antibodies. It is also possible that high

transcript levels of CXCL12 and Notch-1 are present in the BM samples but only in

particular sub-populations of cells. Therefore, this may account for high levels of


expression in the PCR but low numbers of positive cells using flow cytometry, as flow

cytometry did not account for the level of protein present, only the number of

positive cells.

A further explanation to account for differences in gene expression and presence of

protein which has been highly explored in the literature, is that there is only a weak

correlation between the levels of mRNA and protein within mammalian cells (209).

Messenger RNA to protein studies were conducted using cells derived from a variety

of different organisms including mouse, yeast and bacteria using specialised mRNA

and protein quantification techniques. Results were analysed using both the

Spearman rank coefficient and Pearson correlation coefficient statistical tests to

compare mRNA and protein levels at one time by hundreds of genes and

corresponding proteins. On average, a coefficient correlation value of 0.57 was

established which illustrated a correlation of 57% between mRNA and protein levels

in the different cell types (210). Therefore, mRNA and protein values were not

directly correlated. Several factors have been implicated in this finding which

included varying rates of mRNA synthesis and protein production, stability of the

mRNA and protein, in addition to the state of mRNA synthesis (210). Previous studies

found that mRNA was synthesised at a slower rate compared to proteins i.e. cells

were found to produce two copies of mRNA per hour compared to a dozen copies of

the corresponding protein (210). In one study, median translation rate was

approximately 140 proteins per mRNA per hour however, this varied depending on

the function of the proteins of interest; for example, proteins involved in translation

regulation had much slower rates of translation (211). In addition, mRNA degrades

quicker than proteins with a median half-life of 9 hours compared to proteins which

are five times more stable with a median half-life of 46 hours (211). Housekeeping

genes and small proteins were also found to be more stable (211), whereas genes

and proteins for transcription factors and chromatic organisation were significantly

less stable (211). In addition, the state of mRNA production when analysed

influenced protein production. For example, if the cells were producing mRNA for a


particular protein at a steady state, the variability of protein production was high,

whereas if the mRNA was synthesised at varying rates the corresponding protein out-

put correlated well (212). These concepts provide an insight as to why differences in

mRNA and protein expression occur. However, the general concept implies that

mRNA production may be slower and prone to degradation and therefore, this does

not provide a theory as to why the CXCL12 and Notch-1 mRNA expression by the BM

was higher compared to the level of protein. However, as stated above, the rates of

mRNA synthesis and translation may be gene and protein specific and highly variable.

Polymerase chain reaction, flow cytometry and IF were also used to determine the

gene expression and presence of protein for the HSC molecules, ligands and CD138

in the MCs. Taken together, using both gene and protein methodologies, the MCs

expressed all HSC niche molecules as well as the ligands CXCL12, Ang-1 and Jag-1.

However, similar to the BM analysis, there were some discrepancies between the

gene expression and presence of protein in the MCs using both PCR and flow

cytometry. There were no bands present for Tie-2, CXCL12, Ang-1 or Jag-1 by

endpoint PCR in any of the MCs and only marginal expression of CXCL12, Tie-2 and

Ang-1 (which varied depending on the cell type) and UD expression of Jag-1 in all

MCs, by real-time PCR. Whereas, using flow cytometry, a high proportion of 5TGM1-

GFP cells (64%) and a lower number of 5T33-WT and 5T33-GFP cells (32% and 4%,

respectively) were positive for CXCL12, Tie-2 was present at a low frequency only in

the 5T33-WT and 5TGM1-GFP cells (4-5%), and Jag-1 was present only in the 5TGM1-

GFP cells (7%). The 5T33-GFP and 5TGM1-GFP cells were also positive for CXCL12

using IF, where staining seemed to exist as discrete foci and there was no staining of

Jag-1 or Tie-2 in the 5T33-GFP cells and 5TGM1-GFP cells. Again, the positive CXCL12,

Jag-1 and Tie-2 staining in the flow cytometry and/or IF may be due to non-specific

binding of the antibodies and WB could be used to determine the specificity. Low

expression of CXCL12 and Tie-2, and UD expression of Jag-1 by the MCs using

endpoint and real-time PCR could also be due to the inefficiency of the primers and

probes used. However, the primers for endpoint PCR worked successfully using the


BM cDNA and the efficiency of the CXCL12 and Jag-1 probes were within the normal

range as shown in Appendix 2, Section 2.5. The difference in gene expression and

presence of protein may also be due to differences in mRNA and protein production

rates as stated above, which may account for lower mRNA levels of CXCL12, Tie-2

and Jag-1 compared to protein. In addition, sub-populations of cells could also be

responsible. A small sub-population of MCs may express high levels of the molecules

Tie-2 and Jag-1 however, because there are only low numbers of positive cells these

are undetectable using real-time PCR.

Another theory with regards to CXCL12 and the conflicting results between gene

expression and presence of protein in the MCs, is that CXCL12 is internalised from

the FBS in the media that the MCs are cultured in, resulting in false positives using

flow cytometry and IF. In support of this, Hatse et al (213), reported that CXCL12 was

internalised in bovine endothelial cells and in the T lymphoblastic cell line SupT1,

visualised by flow cytometry using a fluorescent recombinant CXCL12 protein.

However, further work is required to test this hypothesis in MCs.

There were also some differences in the expression of some of the molecules and

ligands between the MCs using real-time PCR, flow cytometry and IF. Differences

between the cell types was particularly apparent with regards to N-cad using real-

time PCR, CXCL12, Notch-1 and Jag-1 using flow cytometry and CXCL12 using IF. The

differences observed between the MCs maybe caused by several factors. Firstly,

differences between the 5T33-WT and GFP cells may be due to the selection process

after the GFP had been transduced into the 5T33MMvt cells (the process for this is

unknown as the cells were transduced outside of the University) or their

characteristics may have changed due to in vitro cell culture. The 5TGM1-GFP cell

line, which was derived from the 5T33MMvt cells, demonstrates different

characteristics in vivo, such as bone lesions; therefore, it may be unsurprising that

the expression of some of the molecules differs in vitro.


Flow cytometry was also used to detect the number of N-cad positive 5T33-GFP and

5TGM1-GFP cells. However, due to nonspecific binding of the primary antibody and

unsuitable isotype control antibody, WB was used as an alternative. Western blotting

was originally conducted using two different types of ECL. The ECL purchased from

Thermo Scientific resulted in over-exposed bands despite less exposure time,

whereas ECL purchased from Santa Cruz Biotechnology demonstrated more

optimum band density despite a longer exposure time. Using both types of ECL, the

5T33-GFP lysates did not produce a band for N-cad whereas the 5TGM1-GFP lysates

produced a band of the correct size. Therefore, these data correlated with the real-

time PCR results where N-cad was not expressed in the 5T33-GFP cells but was

detected in the 5TGM1-GFP cells. Western blotting is in some ways more

advantageous than flow cytometry, because in addition to detecting the presence of

protein is can also provide evidence for the specificity of the antibody. However, one

of its limitations is that it is unable to provide a percentage of cells, which are positive

for the protein of interest (POI). Therefore, from these studies it implies that N-cad

is present in the 5TGM1-GFP cells but the percentage of positive cells is unknown.

Taken together, the PCR, flow cytometry and IF data demonstrated that the HSC

niche molecules were expressed by or present on the 5T33-WT, 5T33-GFP and

5TGM1-GFP cells. These data are supported by previous studies, which have

determined the expression of some of these molecules by MCs in vitro. CXCR4

expression was previously determined by the 5T2MM and 5T33MMvt cells (170) and

various human MCs (193-195) in vitro. Notch-1 was expressed by human MC lines

and patient cells (198, 199, 214) and Tie-2 was expressed by the XG-1 cell line in vitro

(201). N-cad was also expressed by several different human cell lines as well as

patient cells (205-207). The MCs were also potentially positive for the ligands CXCL12,

Jag-1 and Ang-1. CXCL12 expression was previously shown in MCs such as ARH-77,

WL2, JIMI, LP-1, NCI-H929 and RPMI-8266 cells using real-time PCR and in human

MM cells using IHC (215). However, as stated previously it is uncertain whether the

MCs in these studies are producing CXCL12 themselves or whether they are


internalising it from the FBS present in the media. Jag-1 protein expression has

previously been demonstrated in primary patient MC samples (198, 214) and Ang-1

expression has also previously been shown in myeloma cell lines including RPMI-

8226, U266, OPM-2, XG-1, and XG-6 as well as patient samples (201).

Endpoint and real-time PCR and flow cytometry demonstrated the expression of the

ligands CXCL12, Jag-1, Ang-1 and N-cad and the receptor Notch-1 by both the MC3T3-

E1 and primary OLCs, whereas Tie-2 was only expressed by the primary OLCs.

However, some of the findings between gene expression and presence of protein did

not entirely correlate. The real-time PCR results implied greater expression of all

ligands with the exception of N-cad by the primary OLCs compared to the MC3T3-E1

cells (though these were not compared statistically). Whereas, using flow cytometry,

the number of Notch-1 and Jag-1 positive cells were higher in the MC3T3-E1 cells

compared to the primary OLCs, and Tie-2 staining was absent by flow cytometry in

the primary OLCs, contradictory to the real-time PCR data. In addition, the difference

in CXCL12 expression between the MC3T3-E1 cells and primary OLCs was obvious

using real-time PCR however, by flow cytometry they differed only by 10 %. Again,

inconsistencies between the gene expression and presence of protein may be due to

differences in transcription and translation rates or the specificity of the PCR and flow

cytometry reactions. The flow cytometry protocol may be particularly responsible for

the contrasting results for Jag-1 and Tie-2. These are both membrane-bound

proteins, and despite cautious efforts to retain its presence through careful

dissociation of the cells from their flasks, the molecules may have been cleaved. This

therefore may have resulted in the absence of Tie-2 and low Jag-1 positive cells in

the primary OLCs using flow cytometry. In addition, flow cytometry only accounted

for the number of cells, which were positive for each molecule, not the number of

proteins present on a per cell basis therefore, this could result in differences between

real-time PCR and flow cytometry.


The overall expression of the HSC niche molecules and ligands by osteoblastic cells

are however, supported by the literature. Jag-1 expression has been detected in

MC3T3-E1 cells (200, 216), BM stromal cells (199), C3H10T1/2 and C2C12 cell lines,

murine primary osteoblastic cells (216) and osteoblasts in mice in vivo (82). Notch-1

expression has previously been identified in MC3T3-E1 cells (200, 216), C3H10T1/2

and C2C12 cell lines and murine primary osteoblastic cells (216). CXCL12 was

expressed in MC3T3-E1 cells, human osteoblastic cells, stromal cells, MG63 cells

(196) and by MSCs (217). Tie-2 is generally perceived as an endothelial marker

however, Jeong et al (202) determined its expression in MC3T3-E1 cells, murine

primary osteoblastic cells, C2C12 cells, C3H10T1/2 cells and ST-stromal cells. Ang-1

expression has previously been identified in murine primary osteoblastic cells, ST-

stromal cells, C2C12 cells and C3H10T1/2 cells (202). N-cad expression has previously

been identified in MC3T3-E1 cells, C3H10T1/2, ST-stromal cells, ATDC5 chondrogenic

cells, NIH3T3 fibroblastic cells (218), KS483 osteoblastic cells (206), SAOS-2 cells (219)

and by both oval and spindle shaped osteoblasts in vivo (83).

In conclusion, these data support the hypothesis that MCs express and produce the

same repertoire of molecules as HSCs, which include CXCR4, Notch-1, Tie-2 and N-

cad and that their complementary ligands; CXCL12, Jag-1, Ang-1 and N-cad are

expressed by osteoblastic cells in vitro. However, their role within a MC niche is still

unknown and therefore, further research is required to explore this.

Chapter IV: The ex vivo and in vivo expression of HSC niche molecules and their ligands by murine myeloma cells Page 131

Chapter IV: The ex vivo and in vivo expression of HSC

niche molecules and their ligands by murine myeloma

cells


4.1 Introduction

As discussed previously, it has been suggested that MCs may home to a BM

microenvironmental niche similar to that occupied by HSCs. To achieve this it was

hypothesised that MCs express the same repertoire of HSC niche molecules, which

anchor MCs to the niche and induce MC dormancy.

Chapter III explored the expression and presence of the HSC niche molecules and

complementary ligands in murine MCs (5T33-WT, 5T33-GFP and 5TGM1-GFP) and

osteoblastic cells (MC3T3-E1 and primary OLCs) cultured in vitro. However, the

expression of these molecules was not determined ex vivo or in vivo. It remains

unknown whether the MCs express/utilise the same HSC niche molecules once

injected into mice and also whether these molecules play a role in the homing,

growth and subsequent survival of the MCs within the BM in vivo.

Previous studies have determined the expression of some of the HSC niche molecules

and their ligands in MCs isolated from BM or peripheral blood of mice or patients

analysed ex vivo or in vivo using BM sections.

Van de Broek et al (220) demonstrated that 59% of MM patients had CXCR4 positive

BM MCs, when analysed ex vivo and this is supported by a separate study which

found that 5/5 myeloma patient samples were positive for CXCR4 when isolated and

analysed in a similar way (221). In addition, Alsayed et al (195) reported that both

BM and peripheral blood MCs from MGUS and MM patients were positive for CXCR4

using flow cytometry, analysed ex vivo.


CXCL12 expression was detected in human BM cells isolated from healthy donors and

specifically in sorted mononuclear cell fractions, analysed ex vivo (193). CXCL12

protein was also detected in human bone by immature osteoblastic cells lining the

endosteum and stromal cells in the marrow space in vivo using IHC (111). In addition,

the presence of CXCL12 was significantly higher in BM samples isolated from patients

with MM compared to healthy controls and there were significantly higher levels of

CXCL12 in the BM compared to samples from the peripheral blood (195).

Previous studies have also demonstrated the expression of Notch-1 and Jag-1 by MCs

and osteoblastic cells ex vivo and in vivo. Patient MCs were isolated from BM samples

by CD38high/CD19low flow cytometry sorting and were cytospun on to slides followed

by IHC staining for both Notch-1 and CD138. CD138 staining was observed on the

membrane of the cells, whereas Notch-1 staining was cytoplasmic (198). The

presence of Notch-1 protein in MCs was also visualised in vivo on BM trephines using

IHC where 72 out of 78 (92.31%) MM patient samples were positive for Notch-1

(214). Jag-1, the ligand for Notch-1 was expressed by human BM stromal cells

isolated from healthy donors and analysed by WB (199). Jag-1 staining was also

present on osteopontin (Opn) positive osteoblasts situated on trabecular bone

surfaces on histological slides using IHC (82).

As stated in Chapter III, there is little evidence in support of Tie-2 expression by MCs.

However, Tie-2 expression was found in 4 out of 23 BM MC patient samples,

determined by RT-PCR. In addition, 11 out of 22 samples also expressed Ang-1.

Furthermore, stromal cells isolated from the BM of myeloma patients were positive

for both Ang-1 and Tie-2 (201). In addition, the presence of Ang-1 protein was also

detected by murine osteocalcin (Ocn) positive osteoblasts lining trabecular bone on

histological slides in vivo (85).


N-cad expression has previously been observed in BM patient MCs as well as patient

serum (205). N-cad was also implicated as a marker of prognosis, where patients with

high circulating N-cad were found to have a poorer prognosis and a greater risk of

death compared to those with low levels. In addition, N-cad protein was present in

both spindle and oval-shaped osteoblasts in vivo (83) and by human primary OLCs

isolated from trabecular bone chips using Northern blotting (152).

Taken together, these studies provide evidence that the HSC niche molecules CXCR4,

Notch-1, Tie-2 and N-cad are expressed by patient MCs isolated from the BM or

circulation and analysed ex vivo, as well as being present on patient MCs analysed in

vivo. In addition, their complementary ligands CXCL12, Jag-1, Ang-1 and N-cad are

expressed by osteoblastic cells isolated from patient BM and analysed ex vivo as well

as osteoblastic cells in vivo.

While the above provides some evidence for the expression and activities of HSC

niche molecules and ligands by both human and murine MCs, there is also little or no

data, which systematically determines the expression or function of these molecules

in the 5T33-GFP and 5TGM1-GFP models of myeloma either analysed ex vivo or in

vivo.

In this chapter, I have determined the presence of protein for CXCR4, Notch-1, Tie-2

and N-cad in BM derived 5T33-GFP cells as well as BM and spleen derived 5TGM1-

GFP cells analysed ex vivo using flow cytometry. In addition, I have also mapped the

growth of the 5TGM1-GFP cells and the presence of the HSC niche molecules and

ligands over-time in vivo.



4.2.1. Hypothesis and aims

The aim of these studies was to test the hypothesis that “Myeloma cells express the

same repertoire of molecules as HSCs when analysed ex vivo and in vivo.

4.2.2 Objectives


1. To determine the presence of protein for the HSC niche molecules; CXCR4,

Notch-1 and Tie-2, complementary ligands; CXCL12 and Jag-1, and CD138 in

the BM derived 5T33-GFP cells as well as BM and spleen derived 5TGM1-GFP

cell lines analysed ex vivo using flow cytometry.

2. To determine the presence of CXCR4 and N-cad protein as well as CD138 by

the 5TGM1-GFP cell lines grown in vivo over-time using IHC.

4.3 Methods

4.3.1 Murine models of multiple myeloma

The 5T33-GFP and the 5TGM1-GFP murine models of MM were used to determine

the presence of protein in BM and spleen derived HSC niche molecules and ligands.

The individual characteristics of each murine myeloma model are described in

Chapter I Section 1.5 and the experimental design for these studies are shown below

in Figure 4.1. The study design for the 5T33-GFP model of MM is shown in Figure

4.1a. Due to the variability of the tumour in this particular model, mice were

sacrificed individually when they became ill and a control PBS treated mouse was

sacrificed at the same time. The longitudinal study design for the 5TGM1-GFP model


of MM is shown in Figure 4.1b. At the final time point (23 days P-I) all mice were

sacrificed at the same time rather than individually because the 5TGM1-GFP model

was intrinsically less variable than the 5T33-GFP model.

Figure 4.1: A schematic diagram of the 5T33-GFP and 5TGM1-GFP models of MM. a. One million 5T33-GFP cells or PBS were injected IV via the tail vein into C57BL/KaLwRij mice and upon presentation of hind leg paralysis 4-6 weeks later, mice were sacrificed and samples were collected for ex vivo analysis of HSC niche molecule and ligand protein (N=8). b. Two million 5TGM1-GFP cells or PBS were injected into C57BL/KaLwRij mice and at 10, 14, 17 and 23 days P-I (when the first mouse presented with hind leg paralysis) the mice were sacrificed (N=4-6) and samples were collected for ex vivo analysis for the presence of HSC molecules and ligands.

4.3.2 Protein analysis of bone marrow and spleen derived myeloma cells, analysed

ex vivo

4.3.2.1 Flow cytometry

Murine BM was isolated from the 5T33-GFP, 5TGM1-GFP and PBS injected mice, as

described previously in Chapter II Section 2.2.1.4. In addition, splenocytes were also

isolated from the mice injected with the 5TGM1-GFP cells and complementary PBS

controls, as mice in the 5TGM1-GFP model develop splenomegaly due to infiltration

of tumour cells at this site of haematopoiesis. The spleen was also analysed to

a

b


determine any differences in the presence HSC niche molecules and ligands between

BM and spleen derived cells. The presence of these molecules on either of the cell

types may highlight which molecules may or may not be important for tumour

growth specifically in the BM. These samples were isolated as previously described

in Chapter II Section 2.2.1.4.

Tumour burden was assessed using GFP expression as a marker for the percentage

of MCs in each BM and spleen sample. The samples were also stained with the

optimised quantity of antibody for the HSC niche molecules, ligands and CD138 as

described in Chapter II Section 2.4.1.2. The number of GFP positive MCs, which were

also positive for the HSC niche molecules and ligands, were measured using a FACS

Calibur. Cell Quest software was used to analyse the data by gating against the

negative isotype control antibodies, (which demonstrated less than 1% positive

staining) to produce a percentage of cells, which were positive for each molecule.

Each mouse was classed as a biological replicate to produce a mean percentage of

cells, which were positive for each HSC niche molecule and ligand.

4.3.3 Protein analysis of bone marrow derived myeloma cells, analysed in vivo

4.3.3.1 IHC

Immunohistochemistry was conducted to visualise positive staining of the HSC niche

molecules in the BM cells and MCs in vivo. C57BL/KaLwRij mice were injected with

either 2x106 5TGM1-GFP cells or PBS IV and at 0, 3, 7, 10, 14, 17 and 21 days P-I (N=3-

4 per group) mice were sacrificed as described previously. The hind limbs were

dissected and then processed for histology, as described in Chapter II Section 2.4.3.2.

Wax embedded sections from the 10-21 day time points were stained with anti-

CXCR4, anti-N-cad or anti-CD138 antibodies or complementary isotypes to each

primary antibody. Images of the staining were taken using an Aperio Scan Scope slide

scanner and visualised using Image Scope software. These time points were used, as

clear MC colonies were visible by eye.

CD138 staining by IHC was also analysed by Dr Shelly Lawson (Department of

Oncology, The University of Sheffield) using Osteomeasure Software. The number of

single cells, colony number and area were calculated after 0, 3, 7, 10, 14, 17 and 21

days P-I of 5TGM1-GFP cells (N=3-4 per group).

4.3.4 Statistics

Where relevant, non-parametric statistical tests were used. A Kruskal-Wallis test

followed by Dunn’s post test (P<0.05) was used to determine any differences in

tumour burden and HSC niche molecule, ligand and CD138 protein frequency in the

BM and spleen over-time. A Mann-Whitney U test (P<0.05) was used to compare

tumour burden in the BM of 5T33-GFP injected mice with 5TGM1-GFP injected mice

at the end stage of disease. It was also used to compare the presence of HSC niche

molecules, ligands and CD138 between the 5T33-GFP and 5TGM1-GFP cells at the

end-stage of disease as well as between cells cultured in vitro. Where relevant, error

bars were displayed on all graphs using the SEM, unless stated otherwise in the

legend.

4.4 Results

4.4.1 Quantification of tumour burden at the end-stage of disease in the bone

marrow of mice injected with 5T33-GFP and 5TGM1-GFP cells

Flow cytometry was used to determine the number of 5T33-GFP and 5TGM1-GFP

cells present in the BM of C57BL/KaLwRij mice based upon GFP expression. Figure

4.2a shows an example of the flow cytometry profiles used to quantify the number

of GFP positive MCs in both the 5T33-GFP and 5TGM1-GFP injected mice at the end-

stage of disease. The 5TGM1-GFP injected mice had significantly greater numbers of

GFP positive tumour cells (53.71% ±4.70) (P<0.001) present in their BM compared to

the 5T33-GFP injected mice (2.40% ±0.90). In addition, three of the 5T33-GFP injected

mice included within the mean did not have obvious tumour burden from the flow

cytometry profiles with values less than 0.25% tumour burden. This contrasted with

the mice injected with the 5TGM1-GFP cells, where all mice injected were positive

for GFP positive tumour burden.

Figure 4.2: Green fluorescent protein positive tumour burden was greater in mice injected with 5TGM1-GFP cells compared to 5T33-GFP cells at the end-stage of disease using flow cytometry. One million 5T33-GFP and 2x106 5TGM1-GFP MCs were injected (IV) into 6-8 week old female C57BL/KaLwRij mice. Upon presentation of hind limb paralysis or general illness, the mice were sacrificed (4-6 weeks P-I for the 5T33-GFP injected mice and 23 days for the 5TGM1-GFP injected mice) and the BM was isolated. Tumour burden was assessed by quantifying the number of GFP positive cells using a FACS Calibur and Cell Quest software. a. An example of a representative scatter plot for number of GFP positive cells in the BM samples of 5T33-GFP (i) and 5TGM1-GFP (ii) injected mice. b. Quantitative analysis of the percentage of viable GFP positive cells in the 5T33-GFP and 5TGM1-GFP injected mice. Data was analysed statistically using a Mann Whitney-U test (P<0.05). Results were

determined as a mean percentage of GFP positive cells in the BM of 5T33-GFP injected mice (N=8) and the 5TGM1-GFP injected mice (N=6).

Flow cytometry was also used to determine the percentage of MCs present in the

BM and spleen of mice injected with 5TGM1-GFP cells at 10, 14, 17 and 23 days P-I

as shown in Figure 4.3. From the data, it was clear that there was a linear increase in

GFP positive tumour burden in both the BM and spleen over-time. Statistically, there

was significantly greater tumour burden at 23 days compared to 10 and 14 days in

both the BM and spleen samples (P<0.05). In addition, there seemed to be higher

tumour burden in the BM compared to the spleen, particularly at 14, 17 and 23 days

P-I.

b

ai ii

Figure 4.3: Green fluorescent protein positive tumour burden in the bone marrow and spleen of mice injected with 5TGM1-GFP cells increased over-time and was greater in the bone marrow compared to the spleen using flow cytometry. Two million 5TGM1-GFP cells were injected (IV) into 6-8 week old female C57BL/KaLwRij mice. At 10, 14, 17 and 23 days P-I, the mice were sacrificed and the BM and spleen were isolated. Tumour burden was assessed by quantifying the number of viable GFP positive cells in the BM and spleen of 5TGM1-GFP injected mice using a FACS Calibur and Cell Quest software and displayed as a mean percentage of tumour burden. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s correct test to compare the tumour burden in the BM and spleen separately over-time (P<0.05).

4.4.2 Protein quantification of the HSC niche molecules, ligands and CD138 by 5T33-

GFP and 5TGM1-GFP cells ex vivo using flow cytometry

Flow cytometric analysis was used to determine the number of BM derived 5T33-GFP

and 5TGM1-GFP MCs that were positive for the HSC niche molecules, ligands and

CD138, analysed ex vivo. Table 4.1 shows a summary of the percentage of BM derived

5T33-GFP and 5TGM1-GFP cells, which were positive for each HSC niche molecule,

ligand and CD138, analysed ex vivo. Figure 4.4 shows the mean percentage of 5T33-

GFP and 5TGM1-GFP cells, which were positive for each HSC niche molecule, ligand

and CD138 at the end stage of disease, analysed ex vivo using flow cytometry. A high

percentage of BM derived 5T33-GFP and 5TGM1-GFP cells were positive for the HSC

niche molecules CXCR4, Notch-1 but <2% were positive for Tie-2. The 5T33-GFP cells

were also positive for ligand CXCL12, which was only expressed at a low frequency in

the 5TGM1-GFP cells. Both cell types were also marginally positive for the ligand Jag-

1 and highly positive for the plasma cell marker CD138. Statistically, a significantly

greater number of 5T33-GFP cells were positive for HSC niche molecules CXCR4 and

Notch-1 (P<0.05) and ligands CXCL12 (P<0.05) and Jag-1 (P<0.05) compared to the

5TGM1-GFP cells. There was however, no statistical difference in the number of

5T33-GFP and 5TGM1-GFP cells, which were positive for Tie-2 or CD138.

Table 4.1: A summary of the quantitative analysis for the percentage of bone marrow derived myeloma cells, which were positive for the HSC niche molecules, ligands and CD138 using flow cytometry

Protein 5T33-GFP 5TGM1-GFP

CXCR4 99.53 ± 0.12 93.63 ± 2.05 Notch-1 99.09 ± 0.27 86.83 ± 1.54

Tie-2 1.91 ± 0.23 1.66 ± 0.62 CXCL12 38.18 ± 3.03 1.01 ± 0.39

Jag-1 8.55 ± 0.93 1.88 ± 0.37 CD138 93.28 ± 1.78 93.19 ± 0.70

Mean percentage of positive cells shown using the SEM.

Figure 4.4: 5T33-GFP and 5TGM1-GFP cells isolated from the bone marrow of mice were positive for the HSC niche molecules and ligands at varying frequencies at the end-stage of disease, analysed ex vivo using flow cytometry. One million 5T33-GFP and 2x106 5TGM1-GFP MCs were injected (IV) into 6-8 week old female C57BL/KaLwRij mice. Upon presentation of hind limb paralysis or general illness, the mice were sacrificed (4-6 weeks P-I for the 5T33-GFP injected mice and 23 days for the 5TGM1-GFP injected mice) and the BM was isolated. Cells were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4 (i), Notch-1 (iii) and Tie-2 (v), ligands CXCL12 (ii) and Jag-1 (iv) and CD138 (vi) and dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. Data was analysed statistically using a Mann-Whitney U test (P<0.05). Results were determined as a mean percentage of 5T33-GFP (N=8) and the 5TGM1-GFP (N=6) which were positive for each molecule.

iii

ii

iv

v vi

i

The number of BM derived 5T33-GFP and 5TGM1-GFP cells which were positive for

each HSC niche molecule, ligand and CD138 at the end-stage of disease were

compared to the number of cells which were positive for these molecules in vitro (the

results for this were originally shown in Chapter III, Section 3.4.3). This is displayed in

Figure 4.5 for the 5T33-GFP cells and Figure 4.6 for the 5TGM1-GFP cells.

A significantly greater number of BM derived 5T33-GFP cells (“5T33-GFP ex vivo”)

were positive for the HSC niche molecules CXCR4 (P<0.05), Notch-1 (P<0.05) and Tie-

2 (P<0.05) compared to the 5T33-GFP cells cultured in vitro (“5T33-GFP in vitro”).

There was also a significantly greater number of “5T33-GFP ex vivo” cells which were

positive for the ligands CXCL12 (P<0.05) and Jag-1 (P<0.05) compared to the “5T33 in

vitro” cells.

In contrast, significantly fewer BM derived 5TGM1-GFP cells (“5TGM1-GFP ex vivo”)

were positive for the HSC niche molecules Notch-1 (P<0.05) and Tie-2 (P<0.05)

compared to 5TGM1-GFP cells cultured in vitro (“5TGM1-GFP in vitro”). There was

also a significantly lower number of “5TGM1-GFP ex vivo” cells which were positive

for the ligands CXCL12 (P<0.05) and Jag-1 (P<0.05) compared to the “5TGM1-GFP

cells in vitro” cells. In addition, there was a significant decrease in the number of

“5TGM1-GFP ex vivo” cells which were positive for CD138 compared to the “5TGM1-

GFP in vitro” cells (P<0.05).

Figure 4.5: The presence of the HSC niche molecules and ligands was significantly higher in the bone marrow derived 5T33-GFP cells compared to in vitro 5T33-GFP cells by flow cytometry. 5T33-GFP cells were cultured in vitro (N=3), or in vivo, by injecting 6-8 week old female C57BL/KaLwRij mice with 1x106 5T33-GFP (N=8) cells. At the end-stage of disease (when mice demonstrated hind-leg paralysis), the BM was subsequently isolated for flow cytometry. Both the 5T33-GFP cells analysed ex vivo and in vitro were incubated with directly conjugated primary antibodies for the HSC niche molecules; CXCR4 (i), Notch-1 (iii) and Tie-2 (v), ligands; CXCL12 (ii) and Jag-1 (iv), and CD138 (vi) or dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. Quantitative analysis for the percentage of viable 5T33-GFP cells analysed ex vivo and in vitro which were positive for each HSC niche molecule, ligand and CD138 was determined. Data was analysed statistically using a Mann-Whitney U test (P<0.05) to compare the percentage of 5T33-GFP cells which were positive for each HSC niche molecule, ligand and CD138. Results were determined as a mean percentage of 5T33-GFP cells, which were positive for each molecule.

iii

ii

iv

v vi

i

Figure 4.6: The presence of the HSC niche molecules and ligands was significantly decreased in the BM derived 5TGM1-GFP cells compared to in vitro 5TGM1-GFP cells by flow cytometry.5TGM1-GFP cells were cultured in vitro (N=3), or in vivo, by injecting 6-8 week old female C57BL/KaLwRij mice with 2x106 5TGM1-GFP (N=6) cells. At the end-stage of disease (when mice demonstrated hind-leg paralysis), the BM was subsequently isolated for flow cytometry. Both the 5TGM1-GFP cells analysed ex vivo and in vitro were incubated with directly conjugated primary antibodies for the HSC niche molecules; CXCR4 (i), Notch-1 (iii) and Tie-2 (v), ligands; CXCL12 (ii) and Jag-1 (iv), and CD138 (vi) or dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. Quantitative analysis of the percentage of viable 5TGM1-GFP cells analysed ex vivo and in vitro which were positive for each HSC niche molecule, ligand and CD138 was determined. Data was analysed statistically using a Mann-Whitney U test (P<0.05) to compare the percentage of 5TGM1-GFP cells which were positive for each HSC niche molecule, ligand and CD138. Results were determined as a mean percentage of 5TGM1-GFP cells positive for each molecule.

iii

ii

iv

v vi

i


4.4.3 Protein quantification of the HSC niche molecules and ligands by the bone

marrow and spleen-derived 5TGM1-GFP cells over-time analysed ex vivo using flow

cytometry.

The presence of HSC molecules and ligands in the BM and spleen derived 5TGM1-

GFP cells was also quantified using flow cytometry over-time at 10, 14, 17 and 23

days P-I of tumour cells, as shown in Figure 4.7. The HSC niche molecules CXCR4,

Notch-1 and Tie-2, ligands CXCL12 and Jag-1, and plasma cell marker CD138 were

present in both the BM and spleen derived 5TGM1-GFP cells at each time point.

Flow cytometry showed that there were no clear patterns with regards to the

frequency of HSC niche molecule and ligand positive cells over-time, with the

exception of CXCL12. Using a Kruskal-Wallis test followed by Dunn’s post test there

were significantly fewer BM derived 5TGM1-GFP cells, which were positive for

CXCL12 after 23 days compared to 10 days and interestingly this reduced over the

time course in a linear fashion. This pattern was however, only observed in the BM

derived 5TGM1-GFP cells and not those isolated from the spleen. In addition, the

frequency of Notch-1 positivity increased in the BM derived cells between 10 and 17

days and CD138 was reduced from 14 days to 23 days. The spleen derived 5TGM1-

GFP cells also had significantly reduced numbers of cells, which were positive for

CXCR4 between 10 and 17 days, and a significant increase in CD138 positive cells

from 14 to 23 days.

Differences in the number of BM and spleen derived MCs, which were positive for

the HSC niche molecules and ligands, were unable to be analysed as no suitable non-

parametric two-way tests were available. However, from the data, there did seem to

be fewer BM derived MCs which were positive for CXCL12 compared to those

isolated from the spleen, particularly at 23 days P-I. Interestingly, the number of

Notch-1 positive BM derived cells was lower at 10 days P-I compared to the spleen

derived cells and Tie-2 was absent in the spleen derived MCs at 10 days P-I, whereas

a small proportion of BM derived cells were positive at this time point.

Figure 4.7: Bone marrow and spleen derived 5TGM1-GFP cells were positive for the HSC niche molecules and ligands and the frequency of positive cells for some of the molecules changed over-time. Two million 5TGM1-GFP MCs were injected (IV) into 6-8 week old female C57BL/KaLwRij mice. At 10, 14, 17 and 23 days P-I the mice were sacrificed and the BM and spleen were isolated (N=6). Cells were incubated with directly conjugated primary antibodies for the HSC niche molecules CXCR4 (i), Notch-1 (iii) and Tie-2 (v), ligands CXCL12 (ii) and Jag-1 (iv) and CD138 (vi) and dose matched isotype control antibodies. This was followed by flow cytometry using a FACS Calibur and Cell Quest software. Quantitative analysis for the percentage of viable BM and spleen derived 5TGM1-GFP cells, which were positive for each HSC niche molecule, ligand and CD138 was determined. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s post test (P<0.05) to determine differences in the number of BM and spleen derived 5TGM1-GFP which were positive for each molecule separately over-time. Results were determined as a mean percentage of 5TGM1-GFP cells positive for each molecule.

i ii

iii iv

v vi


4.4.4 Protein visualisation of the HSC niche molecules CXCR4 and N-cad, and CD138

by the 5TGM1-GFP cells in histological bone sections, over-time using IHC

Immunohistochemistry was used to visualise the presence of protein for CXCR4, N-

cad and CD138 in vivo over the disease course in histological bone sections taken

from mice injected with 5TGM1-GFP cells or non-tumour bearing animals. Figure 4.8

illustrates CD138 staining in histological bone sections taken from mice injected with

5TGM1-GFP cells or PBS and stained by IHC at 10, 14, 17 and 21 days P-I. In the tibiae

of PBS injected mice, only single cells were positive for CD138, unlike in the 5TGM1-

GFP injected mice where discrete colonies were observed. 5TGM1-GFP cells were

morphologically different to the normal constituents of the BM as they

demonstrated an enlarged nucleus and some cells were also multi-nuclear. Staining

at 10 days P-I showed small CD138 positive colonies, which were mostly adjacent to

sinusoid structures and as the disease progressed over-time the number of colonies

increased. From 17 to 21 days P-I, the CD138 positive colonies had merged and

therefore, colony number decreased despite colony area increasing. Quantification

of the colony number over-time is shown in Figure 4.9 (analysed by Shelly Lawson,

Department of Oncology, The University of Sheffield). These were taken from bone

sections from the same study however, staining was conducted on the early time

points (0, 3 and 7 days P-I) as well as the latter time points (10, 17 and 21) and

analysed using Osteomeasure software. At the early time points of 0, 3 and 7 days,

single cells were observed which were either normal murine plasma cells or MCs,

however, the two are indistinguishable. From 7 days onwards, some colonies were

observed though they were small and not always distinct. Following 7 days there was

a linear increase in tumour burden as seen previously with the flow cytometry

tumour burden data.


Figure 4.8: CD138 IHC staining increased over-time in the tibiae of 5TGM1-GFP injected mice. PBS (a) or 2x106 5TGM1-GFP cells (b-e) were injected into C57BL/KaLwRij mice and at 10 (b), 14 (c), 17 (d) and 21 (e) days P-I, mice were sacrificed and tibiae were dissected. The Tibiae were fixed, decalcified, processed, embedded and cut into 3µm sections. Bone sections were stained by sandwich IHC using an anti-CD138 primary antibody or complementary isotype control antibody followed by biotinylated secondary antibody, streptavidin HRP and DAB substrate where positive staining is brown. Images of sections were taken using an Aperio Slide Scanner and subsequently visualised using Image Scope Software. At each time point, images were taken of primary antibody staining using a 2X magnification (i) followed by images using a 40X magnification of isotype (ii) and primary antibody (iii) staining of trabecular bone regions. Scale bars represent 900 µM (i) and 50 µM (ii and iii).

b

c

d

e

a

ii iii i


Figure 4.9: The number of CD138 colonies increased over-time before merging together. Bone sections of mouse tibiae, taken from mice injected with 5TGM1-GFP cells after 1, 3, 7, 14, 17 and 21 days P-I, were stained for CD138 by IHC. The staining was analysed using Osteomeasure software to quantify the number of CD138 positive stained colonies per bone. (Sections were stained by Julia Hough, Department of Human Metabolism, The University of Sheffield and analysed by Shelly Lawson, Department of Oncology, The University of Sheffield).

Figure 4.10 shows the positive staining of N-cad in histological bone sections of tibiae

taken from mice injected with PBS or 5TGM1-GFP cells over-time. In the PBS control,

positive staining for N-cad was visualised on what appeared to be osteoblastic cells

lining both cortical and trabecular bone throughout the bone and staining was

particularly intense in the growth plate region. In addition, the staining appeared to

be primarily on the membrane of the cells.

In the bones of mice that had been injected with 5TGM1-GFP cells, staining again was

observed in osteoblastic cells in trabecular and cortical regions. In addition, all MC

colonies, identified by morphology, were also positive for N-cad but staining was not

as intense as the CD138 IHC staining. There was also a direct correlation of tumour

burden with N-cad staining and as the tumour burden increased so did the number

of N-cad stained cells and tumour area. The staining pattern of N-cad also differed


between the osteoblastic and MCs. The osteoblastic staining for N-cad was highly

membranous, whereas the staining of N-cad in the MCs seemed to be cytoplasmic.

In addition, the intensity of the staining seemed to increase until day 14 and following

this, at 17 and 21 days P-I the intensity of staining decreased.

Figure 4.11 illustrates IHC staining of CXCR4 in the BM of PBS or 5TGM1-GFP injected

mice. Very little staining, if any, was present in the PBS injected mice, whereas in the

5TGM1-GFP injected mice the staining was highly selective for the 5TGM1-GFP cells.

In addition, all 5TGM1-GFP cells seemed to be positive for CXCR4 and staining

correlated with an increase in tumour burden over-time. The staining pattern within

the MCs was quite unusual and unexpected. The cells had both cytoplasmic staining

and intense nuclear staining, which is not typical of G-protein coupled receptors

(GPCRs). In addition, the staining intensity seemed to increase as the tumours

increased in size over-time.


Figure 4.10: N-cad IHC staining increased over-time in the tibiae of 5TGM1-GFP injected mice. PBS (a) or 2x106 5TGM1-GFP cells (b-e) were injected into C57BL/KaLwRij mice and at 10 (b), 14 (c), 17 (d) and 21 (e) days P-I, mice were sacrificed and tibiae were dissected. The Tibiae were fixed, decalcified, processed, embedded and cut into 3µm sections. Bone sections were stained by sandwich IHC using an anti-N-cad primary antibody or complementary isotype control antibody followed by biotinylated secondary antibody, streptavidin HRP and DAB substrate where positive staining is brown. Images of sections were taken using an Aperio Slide Scanner and subsequently visualised using Image Scope Software. At each time point, images were taken of primary antibody staining using a 2X magnification (i) followed by images using a 40X magnification of isotype (ii) and primary antibody (iii) staining of trabecular bone regions. Scale bars represent 900 µM (i) and 50 µM (ii and iii).

a

b

c

d

e

ii iii i


Figure 4.11: CXCR4 IHC staining increased over-time in the tibiae of 5TGM1-GFP injected mice. PBS (a) or 2x106 5TGM1-GFP cells (b-e) were injected into C57BL/KaLwRij mice and at 10 (b), 14 (c), 17 (d) and 21 (e) days P-I, mice were sacrificed and tibiae were dissected. The Tibiae were fixed, decalcified, processed, embedded and cut into 3µm sections. Bone sections were stained by sandwich IHC using an anti-CXCR4 primary antibody or complementary isotype control antibody followed by biotinylated secondary antibody, streptavidin HRP and DAB substrate where positive staining is brown. Images of sections were taken using an Aperio Slide Scanner and subsequently visualised using Image Scope Software. At each time point, images were taken of primary antibody staining using a 2X magnification (i) followed by images using a 40X magnification of isotype (ii) and primary antibody (iii) staining of trabecular bone regions. Scale bars represent 900 µM (i) and 50 µM (ii and iii).

a

b

c

d

e

ii iii i


4.5 Discussion

The aim of this chapter was to determine the number of 5T33-GFP and 5TGM1-GFP

cells, isolated from the BM and/or spleen and analysed ex vivo, which were positive

for the HSC niche molecules. An additional aim was to determine the presence of

protein for HSC niche molecules by 5TGM1-GFP cells in histological bone sections at

different stages of disease.

Firstly, the tumour burden in the BM of mice injected with 5T33-GFP or 5TGM1-GFP

cells was determined by GFP expression, using flow cytometry. Despite 5TGM1-GFP

cells being a derivative of the 5T33MMvt cells (174) there were significant differences

in tumour burden. The flow cytometry data demonstrated that the 5TGM1-GFP

injected mice had significantly greater tumour burden in the BM compared to mice

injected with 5T33-GFP cells. In addition, the take rate was 100% in the mice injected

with 5TGM1-GFP cells whereas; the take rate was 63% for mice injected with 5T33-

GFP cells. This was slightly unusual; however, the preparation of the 5TGM1-GFP cells

differs to that of the 5T33-GFP cells. As described by Dallas et. al (222) the 5TGM1-

GFP cells were “propagated by marrow transfer” in the C57BL/KaLwRij mice, by BM

flushes followed by growth in vitro for 7 days and re-inoculation into mice, whereas

the 5T33-GFP cells were continually cultured without in vivo passaging. Therefore,

passaging through the animals may result in highly tumourgenic 5TGM1-GFP cells. It

maybe that if the 5TGM1-GFP cells had been cultured extensively, as the 5T33-GFP

cells had been, they would have lost certain characteristics required for the homing

of cells to the BM and subsequent growth in vivo. Within our group, we have

attempted to increase tumour take by in vivo passaging of the 5T33-GFP cells but

with little success. However, continual in vivo passaging of the 5T33-GFP cells may be

required to recapitulate the original tumourgenicity of these cells.


In the 5TGM1-GFP injected mice, tumour burden was observed in both the BM and

spleen (sites of haematopoiesis in mice after birth), which is not typical of human

myeloma disease. Tumour burden at both sites increased significantly over-time. The

5TGM1-GFP cells grew in a linear fashion with an initial lag phase between 10 and 14

days (the earliest time points where GFP positive cells were clearly visualised using

flow cytometry) followed by an exponential growth phase from 14 days up until 23

days. Interestingly, there was less tumour burden in the spleen at each time point

compared to the BM (though this could not be statistically compared). These data

therefore, suggested that the 5TGM1-GFP cells may either initially home to the BM

and subsequently migrate to the spleen, or that the MCs arrive in the BM and spleen

but grow at a slower rate in the spleen compared to the BM. Oyajobi et al (223)

observed a high number of 5TGM1-GFP cells in the BM of live mice (Light Tools

Imaging system) whereas, only a low number of 5TGM1-GFP cells were seen in the

spleen of animals 4 weeks P-I. However, again, it is uncertain whether fewer cells

home to the spleen initially or whether they proliferate at a slower rate compared to

those in the BM. Additional studies are therefore, required to ascertain the number

of cells initially homing to both the BM and spleen.

The presence of HSC niche molecules; CXCR4, Notch-1 and Tie-2, their respective

ligands; CXCL12 and Jag-1, and CD138 were shown in 5T33-GFP and 5TGM1-GFP cells

isolated from the BM and analysed ex vivo. When the number of cells which were

positive for the HSC niche molecules, ligands and CD138 were compared between

the 5T33-GFP and 5TGM1-GFP cells isolated at the end stage of disease, several

differences were identified. Significantly less BM derived 5TGM1-GFP cells were

positive for CXCR4, Notch-1, CXCL12 and Jag-1 compared to the 5T33-GFP cells.

Currently there is no data available that have compared the expression of these

molecules between the two different MC lines. In my studies, differences between

the cell lines at the time point when the 5TGM1-GFP treated animals were at end-

stage disease may be due to the differences in tumour infiltration in the BM. These

molecules may be important in the early establishment of disease and therefore are


more highly expressed in the 5T33-GFP cells compared to the 5TGM1-GFP cells,

which had fully infiltrated the BM at this time of analysis. In previous studies, CXCR4,

Notch and Tie-2 signalling have been associated with several different mechanisms

in healthy cells as well as cancer cells. Specifically, CXCR4 is important for chemotaxis

via the activation of PI3-K (105) and MAPK/ERK (103) signalling following CXCL12

stimulation of CXCR4. Notch-1 signalling also induces cell growth and enhances cell

survival potentially via ERK and Akt signalling, following Notch-1 activation (224).

Phosphatidylinositol 3-kinase and Akt signalling have also been associated with cell

survival following activation of Tie-2 by Ang-1 (85). Chemotaxis, migration and

proliferation are often mechanisms, which are important in the early time points of

disease establishment therefore, it is likely that the molecules CXCR4, Tie-2, and

Notch-1 which induce chemotaxis and/or proliferation are down-regulated in the

5TGM1-GFP cells, which have successfully homed and established full disease in the

BM compared to the 5T33-GFP cells.

Previous studies have supported as well as disapproved the theory that early disease

MCs have higher expression of the molecules CXCR4, Notch-1, Tie-2, CXCL12 and Jag-

1 compared to MCs in established disease. With regards to Notch-1, Skrtic et al (2010)

found that MGUS patients, with less than 10% tumour infiltration, had MCs which

were Notch-1 negative, compared with symptomatic myeloma patients, who had

high Notch-1 expression (214). This is similar to prostate cancer where 41% of

patients with high-grade disease and 61% of patients with prostate metastasis had

Notch-1 positive prostate cancer cells, compared with 23% of patient with low-

graded prostate cancer (225). Taken together, this shows that Notch-1 seems to be

present in latter stages of disease. The presence of Jag-1 on CD138 positive MCs has

previously been demonstrated in primary patient samples (198) but previous studies

have not identified its expression on murine 5T cells in vitro, ex vivo or in vivo. Jag-1

expression seems to be associated with disease progression in myeloma. Samples

from MGUS patients were negative for Jag-1 staining using IHC whereas, samples

from symptomatic myeloma patients were positively stained for Jag-1 (214). This was


also demonstrated in prostate cancer, where localised primary tumours and

metastatic samples had higher intensity of Jag-1 staining compared to benign

samples. High levels of Jag-1 also correlated with reoccurrence of disease assessed

by prostate-specific antigen free survival (226). Taken together, this disapproves my

findings of higher expression in the early stages of myeloma disease. However, when

comparing the presence of protein for the molecules in the 5TGM1-GFP time-course

experiment rather than between the 5T33-GFP and 5TGM1-GFP cells at the end-

stage of disease the data, Notch-1 correlates with the experiments described above

where Notch-1 expression was significantly lower at 10 days compared to 14 days.

It is also difficult to rationalise why the number of cells positive for CXCL12 were

lower in the 5TGM1-GFP cells compared to the 5T33-GFP cells and also why the

number of CXCL12 positive cells decreased linearly over-time in the 5TGM1-GFP cells

grown in the BM. As stated in Chapter III, whether CXCL12 is produced by the

5T33MMvt and 5TGM1-GFP cells themselves or whether it is internalised from the

local environment is uncertain. However, our hypothesis is that fewer 5TGM1-GFP

cells were positive for CXCL12 towards the end-stage of disease because they no

longer internalised CXCL12 from the BM environment. This may be because CXCL12

levels were reduced in the BM, as fewer CXCL12 producing cells were present due to

the physical expansion of MCs and reduction in CXCL12 positive osteoblastic cells.

The 5T33-GFP cells isolated from the BM however, had a higher frequency of CXCL12

positive cells than those cultured in vitro. This increase maybe due to induced

production by the MCs themselves or because of the internalisation of CXCL12, which

is still in abundance in the BM, due to low 5T33-GFP tumour burden. In support of

CXCL12 internalisation, previous studies demonstrated that CXCL12 was internalised

via CXCR4 in human and mouse ECs (93). In addition, previous studies showed

internalisation of the receptor CXCR4 in MCs (195) and B lymphoblastic cells (227),

which in turn may result in ligand internalisation, however these studies did not

address that question.


The presence of protein for the HSC niche molecules, ligands and CD138 by the 5T33-

GFP and 5TGM1-GFP cells were compared between the BM derived MCs at the end-

stages of disease and the MCs cultured in vitro. In the 5T33-GFP cells, the HSC niche

molecules; CXCR4, Notch-1 and Tie-2 and ligands; CXCL12 and Jag-1 were significantly

up regulated in the BM derived cells compared to the 5T33-GFP cells cultured in vitro.

Whereas, the HSC niche molecules; Notch-1 and Tie-2, ligands; CXCL12 and Jag-1, and

CD138 were significantly down regulated in the BM derived 5TGM1-GFP cells

compared to the 5TGM1-GFP cells cultured in vitro. The data were almost ‘mirror

images’ of each other but again may reflect the difference in tumour growth rates in

each cell line observed in my studies. As with the differences in protein frequency

between the 5T33-GFP and 5TGM1-GFP, exactly why cells cultured in vitro have

different numbers of positive cells to ex vivo cells is unknown. However, cell

localisation in the BM to specific ligands or regulation by secreted factors would be

likely to influence gene expression and protein production. Increases in the HSC

molecules and ligands in the ex vivo 5T33-GFP cells but decreases in the 5TGM1-GFP

cells again may be due to differences in the phase of cell growth. Whilst the 5T33-

GFP cells have not established full tumour infiltration, they produce proteins that

promote their growth whereas, the 5TGM1-GFP cells, which have already reached

full capacity in the BM, down-regulate molecules, which they no longer require, as

discussed above.

In addition, there were also several differences between the BM and spleen derived

5TGM1-GFP cells. Notch-1 was present in a greater number of spleen derived

5TGM1-GFP cells compared to the BM at 10 days P-I and also the number of cells

positive for Tie-2 was greater in the BM compared to the spleen-derived cells at 10

days. Differences in the frequency of these molecules between BM and spleen

derived 5TGM1-GFP cells have not been reported in the literature. However, again

the stage of tumour progression may be important with regards to presence of the

HSC niche molecules and ligands and may account for differences for the presence

of Notch-1. The numbers of cells, which were positive for Tie-2, compared to the BM


and spleen was extremely interesting. At day 10 P-I, the spleen derived 5TGM1-GFP

cells were negative for Tie-2, and it was not until day 17 that they showed similar

numbers of Tie-2 positive cells compared to the BM derived cells. Tie-2 expression

by the 5TGM1-GFP cells remains ambiguous due the lack of gene expression using

RT-PCR. However, it remains interesting that Tie-2 positive cells were observed in the

BM derived cells but were absent initially in spleen derived cells, indicating that Tie-

2 may be a molecule required for initial tumour development specifically within the

BM, opposed than the spleen.

The presence of CXCR4, N-cad and CD138 protein was visualised by 5TGM1-GFP cells

in vivo by IHC of BM sections taken from mice injected with 5TGM1-GFP and

sacrificed at various time points. Using IHC and Osteomeasure analysis, CD138

protein was mapped over-time in the 5TGM1-GFP infiltrated BM sections. The

analysis showed that the tumour burden results correlated perfectly with the flow

cytometry tumour burden data, whereby tumour increased exponentially over-time,

therefore, highlighting the reliability of the staining.

The distribution of N-cad was also mapped using IHC and the staining showed that

all MC colonies were positive for N-cad in vivo at each time-point and the intensity

of N-cad staining reached its peak at 14 days P-I of 5TGM1-GFP cells. The pattern of

staining was also interesting. The osteoblastic cells in the bone had very distinct

membranous staining around the cell whereas; the N-cad positive MCs displayed

cytoplasmic staining. The staining patterns observed correlated partially with the

literature where OPM-1 MCs positive for N-cad demonstrated cytoplasmic staining,

however, they also demonstrated membranous staining (206). This also correlates

with staining observed in patient prostate biopsies where again N-cad staining was

observed in the cytoplasm but also on the membrane (228). However, Zhang et al

(2003) also showed N-cad staining that “asymmetrically localised to the cell surface”

of HSCs and N-cad positive spindle-shaped osteoblasts, rather than inside the cell.


CXCR4 appeared to be present throughout the development of disease however; the

staining patterns were not as expected. From the IF experiments in the Chapter III,

CXCR4 staining was displayed as discrete membranous staining around the cell

however, in the IHC sections, CXCR4 was situated within the nucleus of the cells.

Differences may be due to sample preparation, as the bones were fixed, processed

and the bones were cut into wax sections. This process eliminates the requirement

of permeablisation for antibody entry whereas, the cells for IF were not permeablised

and the CXCR4 antibodies would therefore be unable to enter the cell. Previous

studies also showed that CXCR4 nuclear localisation arises due to stimulation using

CXCL12 in vitro, where treatment of MCs or lymphoblastic cells with recombinant

CXCL12 resulted in internalisation of CXCR4 (195, 227). In addition, previous studies

have demonstrated that nuclear staining of CXCR4 may be a marker for prognosis

however; this remains to be a controversial concept. CXCR4 nuclear IHC staining has

been reported in cancer cells in colo-rectal cancer (229), gall bladder cancer (230)

and interestingly solely in prostate cells in the bone compared to the primary site and

lymph nodes (104) and it is associated with poorer patient outcome. However, Eaton

et al, (231) noted that nuclear CXCR4 staining was removed when prostate samples

were blocked with normal horse serum (presumably the serum block for the

secondary antibody) opposed to casein. The different serum block resulted in fewer

cells, which were positive, and cells that were positive did not display nuclear staining

but demonstrated “cytoplasmic, membranous or punctate positive staining”.

Therefore, this study implies that IHC methodology may result in false positives for

CXCR4 (231). In my studies, the specificity of the antibody needs to be further

evaluated. The antibody used seemed highly specific for the MCs, which express

CXCR4 however, whether it is binding to the correct epitope is unknown. Therefore,

further experiments are required to confirm that the antibody is specific, possibly by

WB. Also further in vitro experiments to explore the effect of CXCL12 stimulation

upon the localisation of CXCR4 in the 5TGM1-GFP cells could be conducted to

determine whether CXCL12 is responsible for CXCR4 internalisation in these cell

types.


In conclusion, these data support the hypothesis that BM derived MCs are positive

for the same repertoire of molecules as HSCs when analysed ex vivo and in vivo.

However, their role in a MC niche is still unknown and further functional studies are

required to explore this.

Chapter V: Validating a model for primary osteoblast lineage cell differentiation in vitro, to determine the expression of HSC niche molecules and ligands, and to develop an osteoblastic and myeloma cell co-culture system Page 162

Chapter V: Validating a model for primary osteoblast

lineage cell differentiation in vitro, to determine the

expression of HSC niche molecules and ligands, and to

develop an osteoblastic and myeloma cell co-culture

system


5.1 Introduction

Osteoblasts are the cells within the bone responsible for production of osteoid,

primarily composed of type 1 collagen which forms mineralised bone when calcified

with hydroxyapatite, Ca10(PO4)6(OH)2 .They are also one of the important

components in the BM microenvironment, which are potentially required to

establish a niche for HSCs, inducing their quiescence. Increases in osteoblast

numbers by constitutively activating the PTH/PTHrP receptor (PPR), using the Col1-

caPPR mouse model, also increased numbers of c-kit+, sca-1+, lin- (KSL) HSCs in vivo

and BM cells from transgenic mice had higher engraftment into irradiated recipients

compared to wild-type control BM cells (82). In addition, inactivation of BMP

receptor 1 A (BMPR1A) in transgenic mice, resulted in an increase in BM osteoblastic

cells and ectopic bone which correlated with an increase in HSCs (83). However,

currently it is unknown whether all cells of the osteoblast cell lineage are involved in

the osteoblastic niche or whether there is a particular sub-set of osteoblasts, which

are particularly important. Zhang et al (83) noted that HSCs seemed to selectively

adhere to a sub-population of N-cad positive, spindle-shaped cells on both trabecular

and cortical bone, clarified to be early osteoblastic cells by Xie et al (84), supported

by Clark et al (232) who defined spindle-shaped osteoblastic cells as immature

osteoblastic cells.

Osteoblasts originate from MSCs through a series of differentiation steps as shown

in Figure 5.1, determined by Friedenstein et al (233).


Figure 5.1: Mesenchymal stem cell differentiation. Mesenchymal stem cells differentiate through a series of different lineages to produce myocytes, tendocytes, adipocytes, chondrocytes, stromal cells and osteoblastic cells. In particular, mesenchymal stem cells form preosteoblasts, followed by mature osteoid producing osteoblasts, which differentiate in to quiescent bone lining cells or osteocytes. Adapted from Harada and Rodan (234) and Caplan and Bruder (235).

Several transcription factors and cellular pathways have been associated with

osteoblast differentiation. Transcription factors of particular importance are runt-

related transcription factor-2 (Runx-2) also known as CBF1α and osterix (Ostx).

Transgenic mice with a deletion for the Runx-2 gene died shortly after birth and

showed signs of dwarfism and an absence of bone formation. Very few osteoblasts

were observed in the BM of embryo sections and those present had a flattened

appearance with only marginal expression of the osteoblast marker osteopontin

(Opn) and an absence of osteocalcin (Ocn) expression (a marker for late stage

differentiation) by in situ hybridisation (78). This was confirmed also by Otto et al

(236). Runx-2 also binds to the Ocn promoter, which stimulates late osteoblast


differentiation, and Runx-2 regulation also correlates with the expression of the bone

differentiation markers; Opn, bone sialoprotein and TGF-β receptor I (237).

Osterix is an important transcription factor for osteoblast differentiation. Over-

expression of Ostx in osteoblast cell lines C2C12 and C3H10T1/2 resulted in an

increase in collagen type I α I (COL1A1) and Ocn expression. Whereas, when the Ostx

gene was deleted in transgenic mice, the mice died postnatally and the embryos had

a complete lack of mineralisation in the skeleton by alizarin red staining. In addition,

Ostx null mice had reduced RNA expression of the early differentiation marker

COL1A1, as well as undetectable bone sialoprotein, osteonectin (On), Opn and Ocn

expression. Therefore, demonstrating that both early and late markers of osteoblast

differentiation were absent within the Ostx null mice. (238).

In addition, several pathways are important for osteoblast differentiation, which

include the Wnt, TGFβ and BMP pathways.

Osteogenic assays have been developed over approximately the last four decades.

Three different primary sources have previously been used (239) including foetal

calvarial cells by enzymatic digestion of calvarial bone (240), BM-derived stromal cells

isolated from iliac crest aspirates or femoral heads during surgery (241, 242) and

trabecular bone explants.

Differentiation has also been mapped over-time and was defined within three stages,

as shown in Figure 5.2 (183, 240), using rat calvarial cells.

The main stages of differentiation are:

1. Proliferation: Genes such as histone genes (H4) (markers of mitosis) are at

maximal levels at 5 days of differentiation. Active proliferation results in high


expression of COL1A1, fibronectin and TGF-β at approximately 10-12 days of

differentiation.

2. Proliferation decline: Proliferation slows from approximately 10 days

onwards and shortly afterwards collagen levels also decline at approximately 15-20

days, after which it plateaus and remains at a basal level for the remainder of

differentiation. Once collagen reaches basal levels, Alp levels reach their maximal at

approximately 20 days.

3. Mineralisation: Upon mineralisation, the Alp levels decline at approximately

28 days but late markers such as Opn, bone sialoprotein and Ocn reach maximal

levels. (183, 240). Detection of mineralisation is commonly assessed by alizarin red

staining, which binds to calcium deposits, and Von Kossa staining, used to visualise

phosphate (239).

Figure 5.2: Stages of osteoblast differentiation. Graph taken from Lian and Stein (183), mapping the stages of osteoblast differentiation using rat calvarial osteoblastic cells and Northern blotting. (Permission from The Iowa Orthopaedic Journal Editor, 27.11.13).

To induce osteoblast differentiation, human and rodent MSCs are generally cultured

with osteogenic media (OGM) containing 10 mM of β-glycerophosphate (BGP) and


50 µg/ ml of ascorbic acid (asc). Βeta-glycerophosphate is the source of phosphate to

promote mineralisation whereas asc is responsible for collagen type I assembly (239).

The addition of BGP in rat calvarial cultures resulted in the formation of mineral

deposits compared to controls cultured in the absence of BGP (243). Bellows et al

(244) noted that mineralisation increased in a dose dependent manner as the

amount of BGP added to the media was increased.

Treatment of MC3T3-E1 cells with 50 µg/ml of asc alone resulted in an increase in

bone sialoprotein, Ocn, PTH and PTHrP expression, which were not detected in cells

cultured in the absence of asc (245). In addition, when BM MSCs were treated with

OGM containing BGP, dexamethasone (dex) and increasing amounts of asc, it

resulted in an increase in cell proliferation, particularly at 250 µM of asc, and it

increased total phosphate and calcium staining particularly at 50 µM asc compared

to controls (241). ST-2 cells cultured in asc alone had markedly increased Alp activity

compared to those treated only with BGP, reaching maximal levels at 18 days.

Ascorbic acid treatment using the ST-2 cells also resulted in increased Ocn mRNA

expression at day 20 and calcium deposits which reached maximal levels at 75 µg/

ml after 24 days (246).

Factors such as glucocorticoids including dex have also previously been used to

enhance differentiation. Human MSCs treated with a combination of dex, asc and

BGP were highly Alp positive at day 8 of culture and by day 12 a “bone-like material”

had formed in the dish and mineralisation was evident (242). Alkaline phosphatase

also increased in a dose dependent manner in MSCs, when cultured with BGP, asc

and increasing doses of dex. The optimum concentration of dex was 10 mM for Alp

and 1000 mM for mineralisation (242). In addition, when dex was administered

alone, it resulted in an increase in Alp over-time in BM stromal cells compared to

controls in standard media (247). Dexamethasone treatment also induced early

expression of Runx-2, Ostx, Ocn and bone sialoprotein (248). Continuous combination


treatments of BGP, asc and dex using rat stromal cells isolated from the long bones

also resulted in rapid proliferation in the first 24 hours, an increase in Alp production

between 7-14 days for cells and Ocn expression between 14 and 21 days after

treatment (249).

As described above, extensive research has previously been conducted to define the

stages of osteoblast differentiation using both human and rodent osteoblastic cells.

It is therefore desirable in these experiments to validate an in vitro osteoblast

differentiation model using mouse calvarial cells which will enable us to investigate

the effects of OGM on murine primary OLCs over-time. Specific time points can then

be used to determine the effect of differentiation upon the expression of the HSC

niche molecules and ligands implicated within a HSC niche and to determine whether

MCs selectively adhere to in vitro osteoblastic cells at a particular stage of

differentiation.


5.2.1 Hypothesis and aims

The aim of these studies was to test the following hypotheses:

1. “Treatment of murine primary osteoblast lineage cells (OLCs) with osteogenic

media (OGM) will result in osteoblast differentiation”.

2. “Differentiating OLCs express HSC niche molecules and ligands in vitro”.

3. “Myeloma cells adhere to OLCs at different stages of differentiation in vitro”.

5.2.2 Objectives



1. To determine whether treatment of murine primary OLCs with OGM results

in osteoblast differentiation, tested by assessment of alkaline phosphatase

(Alp) production, mineralisation and expression of differentiation bone

markers by real-time PCR.

2. To determine whether OLCs express the HSC niche molecules; CXCR4, Notch-

1, Tie-2 and N-cad and ligands; CXCL12, Jag-1, Ang-1 over the course of

differentiation in vitro, using real-time PCR.

3. To establish whether differentiating OLCs adhere to 5TGM1-GFP cells at

specified stages of differentiation in vitro.

5.3 Methods

5.3.1 Assessment of primary osteoblast lineage cell differentiation

5.3.1.1 Primary osteoblast lineage cell differentiation assay

Murine primary OLCs were isolated and cultured as described in Chapter II Section

2.1.2.7 and 2.2.1.2. Primary OLCs were initially seeded in quadruplicate at a density

of 6000 cells/cm2 (N=4) and cultured for 3 days before culturing with OGM

containing 10 mM of BGP and 50 µg/ ml of asc in MEMα-nuc (Day 0). The media was

changed every 2-3 days over the course of 28 days.

5.3.1.2 Alkaline phosphatase analysis

Primary OLCs were seeded in quadruplicate into a 96 well plate using 150 µl standard

or OGM. Cells were harvested on day 0, 3, 7, 10, 14, 21 and 28. At these time points,

cells were lysed using 0.1% triton in PBS and Alp was detected using a PNPP reaction,

assessed by spectroscopy using a SpectraMax 5Me plate reader over the course of

90 min and analysed using SoftMax Pro 5.2 software as described in Chapter II,

Section 2.5.1.2. This was followed by DNA concentration determination using the


Invitrogen PicoGreen® reagent followed by fluorescent spectroscopy using the

SpectraMax 5Me plate reader and analysed using SoftMax Pro 5.2 software as

described in Chapter II, Section 2.5.1.3. This was used to normalise Alp production,

the calculations for which are shown in Appendix 2, Section 2.8, Table 2.6. Results

were based upon four independent experiments (N=4).

5.3.1.3 Mineralisation analysis

Primary OLCs were seeded in quadruplicate into a 12 well plate in 1 ml OGM.

Mineralisation was determined on day 7, 14, 21 and 28 by alizarin red staining as

described in Chapter II, Section 2.5.2.1. Plates were scanned using an Epson 4990

scanner and analysed using ImageJ software. Results were based upon four

independent experiments (N=4).

5.3.1.4 Gene expression analysis by real-time PCR

Primary OLCs were seeded in quadruplicate into a T25 flask using 7 ml OGM. Primary

OLCs were harvested at 7, 14, 21 and 28 days for gene expression analysis.

Ribonucleic acid was extracted, cDNA was synthesised and real-time PCR was

conducted (as described in Chapter II, Section 2.3.1.1, 2.3.2.1 and 2.3.5.2) to

determine the expression of differentiation markers (COL1A1 and Runx-2) and the

HSC niche molecules and ligands. Data was analysed using the comparative ΔΔCt

method, by normalisation to β2M. Data was expressed as a fold change of gene

expression using the week 1 data as the comparator. Results were based upon four


5.3.2 Assessment of 5TGM1-GFP adhesion to primary osteoblast lineage cells at

different stages of differentiation

5.3.2.1 Co-culture assessment

Primary OLCs were differentiated for 3, 14 and 28 days in OGM, after which 5x104

5TGM1-GFP cells were seeded onto the murine primary OLCs for 1 and 6 hours. The

number of 5TGM1-GFP cells adhering to the primary OLCs was assessed by

fluorescent microscopy using a Leica AF6000LX microscope, followed by

quantification using Volocity software. Results were based upon four independent

experiments (N=4).

5.3.3 Statistics

Where relevant, non-parametric statistical tests were used. A Friedman test followed

by Dunn’s post test (P<0.05) was used to compare Alp production, DNA content,

mineralisation, gene expression as well as 5TGM1-GFP cell adhesion to osteoblastic

cells. In addition, a Wilcoxon Matched pairs signed rank test (P<0.05) was used to

determine differences in the adhesion of 5TGM1-GFP cells to primary OLCs after 1

and 6 hours. As a non-parametric paired two-way statistical test was unavailable,

differences in the Alp production and DNA content between the primary OLCs

cultured in standard and OGM media over-time was unable to be conducted. Where

relevant, error bars were displayed on all graphs using the SEM, unless stated

otherwise in the legend.

5.4 Results

5.4.1 Alkaline phosphatase production by primary osteoblast lineage cells cultured

in standard or osteogenic media over-time

Figure 5.3 demonstrates the differences in Alp production between primary OLCs

cultured in standard media and OGM over-time. From the graph, it is clear that Alp

production increased in the primary OLCs cultured in standard media as well as OGM

until 14 days, after which Alp production reduced. Using a Friedman test followed by

Dunn’s post test, the changes in Alp production over-time were significant for both

the sets of samples (P<0.05) however, the Dunn’s post test did not identify at which

time points significant differences in Alp were apparent. It was also clear that Alp

production was higher at each time-point in the OGM treated primary OLCs

compared to the primary OLCs.

Figure 5.3: Primary osteoblast lineage cells cultured in either standard or osteogenic media demonstrated an increase in alkaline phosphatase production over-time. Primary OLCs were cultured in standard MEMα-Nuc or OGM containing 10 mM BGP and 50 µg/ ml asc, beginning at day 0 as indicated by the arrow. At 0, 3, 7, 10, 14, 21 and 28 days, primary OLCs were lysed using 0.1% triton in PBS and the lysates were added to a solution containing 1 mg/ml PNPP and 0.2 M tris buffer. The presence of Alp was measured using a SpectraMax 5Me plate reader for 90 min and analysed using SoftMax Pro 5.2 software. Results were normalised against the DNA content providing a unit of Alp per µg of DNA (U of Alp/ µg DNA) (Y-axis) over-time (X-axis). Data was analysed statistically using a Friedman test followed by Dunn’s post test to determine any differences in Alp by the primary OLCs and primary OLCs cultured in OGM over-time (P<0.05). Results were determined from four independent experiments (N=4).

5.4.2 DNA content of primary osteoblast lineage cells cultured in standard or

osteogenic media over-time

Figure 5.4 demonstrates the DNA content of the primary OLC samples cultured in

standard media and OGM. From the graph, it is clear that as time progressed the

OGM

amount of DNA in both the primary OLC samples and primary OLCs cultured in OGM

samples increased. Therefore, as this was used as a surrogate for cell counts, we can

surmise that the number of cells increased over-time. Statistically, this increase was

significant between days 0 and 28 for both the primary OLCs and primary OLCs

cultured in OGM (P<0.05). Interestingly from 7 days onwards the DNA content in the

primary OLCs cultured in OGM was lower compared to the OLCs cultured in standard

media, thus implying that proliferation in these cells may have decreased due to the

addition of BG and Asc. This however, was not assessed statistically due to a lack of

appropriate test.

Figure 5.4: Primary osteoblast lineage cells cultured in osteogenic media demonstrated a reduction in proliferation over 28 days, assessed by DNA quantification. Primary OLCs were cultured in standard MEMα-Nuc or OGM MEMα-Nuc media containing 10 mM BGP and 50 µg/ ml asc, beginning at day 0 as indicated by the arrow. After 0, 3, 7, 10, 14, 21 and 28 days, primary OLCs were lysed using 0.1% triton in PBS and the DNA content was measured using the PicoGreen® reagent. The DNA present was measured by fluorescence using a SpectraMax 5Me plate reader. Results were calculated by extrapolation from a standard curve of known amounts of DNA (ng). Data was analysed statistically using a Friedman test followed by Dunn’s post test (P<0.05) to determine differences in DNA content over-time. Results were determined from four independent experiments (N=4)

OGM

5.4.3 Mineralisation of primary osteoblast lineage cells cultured in osteogenic

media over-time

Mineralisation was quantified in the primary OLCs cultured in OGM over-time, as

shown in Figure 5.5. Figure 5.5a demonstrates representative images of

mineralisation, where mineralisation was visualised as small red nodules stained

positive with alizarin red. Figure 5.5b shows the percentage of mineralisation,

demonstrating a significant increase in the percentage of mineralisation between 7

and 21 days (P<0.01) of OGM.

Figure 5.5: Primary osteoblast lineage cells cultured in osteogenic media showed an increase in mineralisation until 21 days. Primary OLCs were cultured in OGM MEMα-Nuc media containing 10 mM BGP and 50 µg/ ml asc. At 7, 14, 21 and 28 days of growth in OGM, primary OLCs were fixed in 100% IMS over-night at 4°C followed by staining with 40 mM alizarin red for 90 min at room temperature and washed in 95% IMS and dried over-night. a. Representative images of mineralisation at 7 (i), 14 (ii), 21 (iii) and 28 (iv) days after growth in OGM, captured using an Epson 4990 scanner. b. The Percentage of osteoblast mineralisation was calculated using ImageJ software. Data was analysed statistically using a Friedman test followed by Dunn’s post test (P<0.05). Results were determined from four independent experiments (N=4).

ai ii iii

.

iv

b


5.4.4 Gene expression of osteoblast differentiation markers and HSC niche

molecules and ligands by primary osteoblast lineage cells cultured in osteogenic

media over-time

Table 5.1 shows the Ct values generated using real-time PCR to determine the

expression of the osteoblast differentiation markers COL1A1 and Runx-2, and HSC

niche molecules and ligands by primary OLCs cultured in OGM.

Table 5.1: A summary of the quantitative analysis for the absolute real-time PCR gene expression of the osteoblast differentiation markers, HSC niche molecules and ligands by primary osteoblast lineage cells cultured in osteogenic media over-time

Gene 7 days 14 days 21 days 28 days

COL1A1 18.11 ± 0.56 18.72 ± 0.50 18.84 ± 0.21 19.11 ± 0.16 Runx-2 24.16 ± 0.26 24.08 ± 0.38 23.80 ± 0.52 23.90 ± 0.45 CXCR4 33.25 ± 1.30 32.31 ± 0.17 31.10 ± 0.68 32.10 ± 0.56

Notch-1 26.12 ± 0.42 26.31 ± 0.52 26.01 ± 0.55 26.15 ± 0.52 Tie-2 28.53 ± 1.13 29.24 ± 1.43 28.59 ± 1.07 29.46 ± 1.25 N-cad 25.26 ± 0.20 25.42 ± 0.47 24.83 ± 0.25 25.97 ± 0.61

CXCL12 21.51 ± 1.08 22.55 ± 0.83 22.06 ± 0.98 21.52 ± 1.09 Jag-1 25.02 ± 0.25 25.93 ± 0.54 25.35 ± 0.35 25.95 ± 0.42 Ang-1 26.04 ± 0.42 26.74 ± 0.78 26.11 ± 0.47 25.95 ± 0.42

Mean Ct values ± SEM

Figure 5.6 demonstrates the results graphically using the fold change and percentage

of gene expression above the columns using the comparative ΔΔCt method, using

Day 7 as the comparator. CXCR4 was not shown graphically as the ΔΔCt method of

analysis could not be used, as some of the Ct values were undetermined (UD). The

differentiation markers COL1A1 and Runx-2 were both highly expressed by the

primary OLCs at each time point and COL1A1 had the greatest expression. Over-time,

COL1A1 expression declined, although this change did not reach significance,

whereas, Runx-2 expression was stable across the time course. All of the HSC niche

molecules were also expressed by the OLCs, though CXCR4 expression was

inconsistent across the time course, as some samples were UD. Notch-1 and N-cad

were highly expressed, whereas Tie-2 and CXCR4 were expressed at lower levels. The

ligands CXCL12, Jag-1 and Ang-1 were also expressed by the primary OLCs at each


time point and CXCL12 was expressed at a higher level compared to both Jag-1 and

Ang-1. Statistically, there were no significant differences in the expression of the HSC

niche molecules or ligands over-time in the primary OLCs cultured with OGM. In

addition, no apparent trends in the gene expression of the HSC niche molecules and

ligands were observed over the time course with the exception of Jag-1 and N-cad,

which decreased over-time.


Figure 5.6: There were no significant changes in the gene expression of bone and HSC niche markers over-time by primary osteoblast lineage cells cultured in osteogenic media over 28 days. Primary OLCs were cultured in OGM containing 10 mM BGP and 50 µg/ ml asc. At 7, 14, 21 and 28 days, RNA was extracted, cDNA was synthesised and real-time PCR reactions were run using approximately 40 ng (2 µl) of RT+ cDNA samples. An equal volume of RT- and H2O control samples were used to assess genomic and reagent contamination. Real-time PCR reactions were run using an Applied Biosystems’ 7900HT Real-Time PCR machine and Applied Biosystems’ recommended settings. Data was analysed using the comparative ΔΔCt method and shown graphically using the fold change, and the percentage of gene expression is shown above the columns for collagen-1 α (a), Runx-2 (b) CXCR4 (c), CXCL12 (d), Notch-1 (e), Tie-2 (f), Ang-1 (g) and N-cad (h). Data was analysed statistically using a Friedman test followed by Dunn’s post test. Results were determined from four independent experiments (N=4).

a b

.

c d

f e

h g

5.4.5 Adhesion of 5TGM1-GFP cells to primary osteoblast lineage cells cultured in

osteogenic media

Figure 5.7a demonstrates representative images showing the adhesion of 5TGM1-

GFP cells to osteoblastic cells at the different time-points. Figure 5.7b shows the

differences in the number of 5TGM1-GFP cells adhering to primary OLCs cultured in

OGM for 7, 14 and 28 days after 1 and 6 hours. From the data, there seemed to be

an increase in the number of 5TGM1-GFP cells adhering to the primary OLCs at 6

hours compared to 1 hour at each of the differentiation time points however, using

a Wilcoxon matched pairs signed rank test this difference did not reach significance.

When the number of 5TGM1-GFP cells adhering to the primary OLCs cultured in OGM

over-time were compared there was a decrease in the number of 5TGM1-GFP cells

adhering to primary OLCs cultured in OGM for 28 days compared to 3 days however,

this was only significant at the 6 hour time point (P<0.05).

Figure 5.7: A greater number of 5TGM1-GFP cells adhered to primary osteoblast lineage cells cultured in osteogenic media for 3 days compared to 28 days. Fifty-thousand 5TGM1-GFP MCs were seeded on to primary OLCs cultured in OGM containing 10 mM BGP and 50 µg/ ml asc) for 3, 14 and 28 days. Adhesion after 1 and 6 hours was imaged using a Leica AF6000LX microscope at a 10X magnification using LAS LF software and analysed using Volocity software to determine the number of MCs adhering to primary OLCs per tile by GFP expression. a. Representative images of 5TGM1-GFP cells adhering to primary OLCs cultured in OGM for 3, 14 and 28 days, and after 1 and 6 hours of adhesion. Scale bars represent 100 µM. b. The number of 5TGM1-GFP cells adhering to primary OLCs cultured in OGM for 3, 14 and 28 days after 1 and 6 hours. Data was analysed statistically using a Wilcoxon matched pairs signed rank test (P<0.05) to compare differences between the number of 5TGM1-GFP cells adhering to primary OLCs after 1 and 6 hours and Friedman test followed by Dunn’s post test (P<0.05) to compare differences in the

number of 5TGM1-GFP cells adhering to primary OLCs cultured in OGM for 3, 14 and 28 days. Results were determined from four independent experiments (N=4).

a

b


5.5 Discussion

The aim of these experiments was to firstly, validate an in-house model to

differentiate murine primary OLCs in vitro and secondly to use this system to

determine whether changes in HSC niche molecules and ligand expression occurred

at the different stages of differentiation and whether MCs adhered to a specific type

of osteoblast in vitro, at a specific stage of differentiation.

Firstly, Alp levels were established using the PNPP assay. Alkaline phosphatase is an

enzyme, which is bound to the surface of osteoblasts by a phosphoinositol bond and

is also found within the bone matrix. It is thought to act by removing inhibitors of

hydrodroxyapatite crystal growth and also by increasing concentrations of

phosphorus at the site of bone formation, to increase bone mineralisation (232).

Therefore, Alp production increases before mineralisation occurs.

In these experiments, murine primary OLCs were cultured with either standard

MEMα-nuc or OGM. In both groups, there was an increase in Alp production until

day 14 post treatment. However, this was to a lesser extent in the cells cultured in

standard media compared to OGM. The increase in Alp over-time was also significant

based on the Friedman statistical test however; a Dunn’s post test was unable to

identify the specific time points, which were significantly different. This may be due

to the high variability of the results in these experiments. The variability seemed to

be a result of the first two and latter two experiments being conducted using two

separate isolates of calvarial osteoblasts. Therefore, there seems to be high

variability between different batches of isolated osteoblasts and for future studies,

it would be desirable to use cells from the same batch. This variability may be

unsurprising as previous experiments have experienced variability even within


osteoblastic cell lines where sub-clones of MC3T3-E1 cells expressed different

amounts of Alp, Ocn and bone sialoprotein when differentiated (245).

When comparing the pattern of Alp production over-time this varied slightly from

results in the literature. Lian and Stein et al (240) and Lynch et al (250) reported

maximal Alp production at 19-20 days of differentiation using rat calvarial

osteoblasts, whereas, Alp peaked earlier at 14 days in these experiments. However,

the results in this chapter seemed more so similar to differentiation experiments

using rat stromal cells or MSCs, which demonstrated maximal Alp production

between 14-21 days and 8-12 days respectively (242, 249).

It is also difficult to determine whether previous experiments support the raw Alp

values seen in these experiments, as early experiments mainly use Alp staining and

studies that do use PNPP biochemical staining generally normalise to mg of protein

rather than DNA, therefore, they are not directly comparable. However, a recent

paper by Wang et al (187) demonstrated levels of 0.8 Units of Alp production per µg

of DNA after 7 days of culture in standard media using C57BL/6 murine calvarial

osteoblasts, which was slightly lower but similar to the values calculated in the

studies in this Chapter.

The presence of Alp and an increase in Alp over-time in osteoblastic cells cultured in

standard media has also been described in previous experiments. Jaiswal et al (242)

described how MSCs cultured in standard media, which didn’t contain BGP and asc,

produced Alp, though not to the extent of media containing BGP and asc. In addition,

the type of media used also affected basal levels of Alp, where MSCs cultured in

DMEM had the lowest basal levels whereas, cells cultured in MEMα had the highest

(242).


In the experiments in this chapter, the statistical difference in the Alp produced by

the cells cultured in OGM and standard media over-time could not be performed due

to a lack of non-parametric paired two-way test, however from the graphs it is clear

that the OGM cultured cells do produce higher Alp. Therefore, it can be concluded

that the OGM is causing an increase in Alp production, and potentially inducing

osteoblast differentiation.

Cell proliferation was also investigated in primary OLCs cultured with standard media

and OGM. The concentration of DNA increased over-time significantly for both cell

types from day 0 to day 28. However, again as a suitable test was unavailable to

compare the two cell types it was difficult to confirm the differences in proliferation

between cells culture in OGM compared to those cultured in standard media.

However, by the patterns shown in the graphs it seemed that the OGM cultured cells

had a lower DNA content than those cultured in standard media from 10 days

onwards. Therefore, proliferation may be decreased in the OGM cultured cells. This

is supported by studies conducted by Lian and Stein et al (240) who previously

reported an initial increase in proliferation by histone determination in rat calvarial

osteoblasts between days 5-10, which then reduced and plateaued for the remainder

of the differentiation period examined, also reported by Choi et al (251) using

MC3T3-E1 cells. However, experiments using human MSCs differentiated in OGM

containing dex actually showed an increase in proliferation when measured by crystal

violet staining of DNA (242). Therefore, the way in which proliferation is measured

may provide a difference in determining whether proliferation remains active. Lynch

et al (250) used fluorometric analysis of DNA from rat calvarial osteoblasts cultured

in OGM and found an increase in DNA up to 21 days, supporting Jaiswal et al (242).

Interestingly, in the same experiment, histone 4 was decreased from 5 days onwards,

supporting Lian and Stein et al (240). In addition, the amount of FBS used in the

culture may also affect proliferation rates. Previous papers have shown that some

osteoblastic cells such as the cell line SAOS-2 are cultured in FBS at concentrations as


low as 0.5% (252), which slows proliferation and induces Alp production and

mineralisation.

Mineralisation is a later marker of osteoblast differentiation, associated with a down-

regulation of Alp and an up-regulation of Ocn, Opn and On (183, 240). Mineralisation

in the experiments in this chapter occurred from the beginning of OGM treatment

and peaked at 21 days. Therefore, mineralisation occurred earlier than previous

reports, which showed mineralisation in rat calvarial osteoblastic cells and human

osteoblastic cells at approximately 28 days onwards (183, 253). In addition,

mineralisation usually begins when Alp decreases but mineralisation in these

experiments continued to increase as Alp increased until 14 days. In support of my

data, Nefussi et al (254) found that rat calvarial osteoblastic cells cultured in the

presence of BGP demonstrated mineralisation within 14 days. One potential

explanation for early mineralisation by the OLCs in these experiments is due to high

basal levels of Alp present in the cells due to the type of media used and also,

standard media contains phosphate which, combined with Alp, may stimulate

immediate mineralisation (239).

In addition, inconsistencies in the literature as well as between my data and previous

studies may again be due to the variability between samples and the heterogeneous

nature of the cells. In addition, other groups have found that BGP may not be the

best source of phosphate for mineralisation and Gartland et al (252) found that

inorganic phosphate was more suitable. In addition, it also difficult to conclude the

effect of OGM on mineralisation as a standard media without BGP and asc control

group was not used as a comparator in these experiments. The addition of this

control would have provided evidence of basal levels of mineralisation in the

presence of pre-existing phosphate in the media and therefore, the actual level of

mineralisation due to OGM could have been assessed.


Real-time PCR was conducted to determine the gene expression of the

differentiation markers COL1A1 and Runx-2 in primary OLCs cultured in OGM over-

time. There were no significant differences in COL1A1 and Runx-2 expression over-

time although; there was a slight decrease in COL1A1 expression. At each time point,

COL1A1 expression was extremely high implying that the primary OLCs were

secreting high amounts of collagen protein to begin with. Lian and Stein et al (183)

previously described high COL1A1 expression in the early time points of

differentiation, which drastically dropped after 12 days in rat calvarial osteoblasts.

This was supported by Choi et al (251) using MC3T3-E1 cells whereby COL1A1

expression was highest between 10-15 days and reduced significantly after 15 days.

Therefore, these results coincide with the maximal expression of COL1A1 in the

murine primary OLCs observed at 7 and 14 days. However, the drop after the

maximal expression of COL1A1 was not significant but a longer period of

differentiation may be required to detect a further reduction in the future.

Runx-2 is an early marker of differentiation and its expression has been observed in

a variety of primary OLCs at different stages as well as in MSCs. In the experiments in

this chapter, Runx-2 was expressed by the primary OLCs at each time point but there

was also no significant difference between Runx-2 expression over time. This is

consistent with previous studies which found that the Runx-2 isoform II mRNA was

constitutively expressed by the human stromal cell line hMS2-15 when treated with

BGP, asc and dex over a 14 day period, whereas, the isoform I was not expressed at

all (255). Another study determined the expression of Runx-2 in a variety of stromal

and osteoblastic cells, which included “non-osteoblastic”; Cos-7, NIH3T3 and

C3H10T1/2 cells, immature osteoblastic; ST-2 and MC3T3-E1 cells and mature

osteoblastic; Ros 17/2.8, FRC, SaOS-2 and U20S cells. Each cell type expressed the

RNA (assessed by Northern blot) for Runx-2 isoform I and both immature and mature

osteoblastic cells expressed Runx-2 isoform II, whereas the RNA from the non-

osteoblastic cells for Runx-2 isoform II was negative. In contrast, using PCR the mRNA

from the non-osteoblastic cells was positive for Runx-2 isoform II. The protein


expression was also different from the RNA and mRNA results where only the mature

osteoblastic cells and the ROS 17/2.8 cells were positive for each isoform, whereas

the immature osteoblastic cells were negative (256). This therefore, demonstrates

the complexities of using Runx-2 as a marker for osteoblast differentiation and it is

difficult to determine whether Runx-2 expression does differ over the course of

differentiation due to the different isoforms and inconsistencies in RNA, mRNA and

protein expression.

To assess differences in the expression of HSC niche molecules and ligands by OGM

treated primary OLCs, real-time PCR was used. Interestingly, the murine primary

OLCs expressed CXCR4 at each time-point however; this was inconsistent as some

values were UD between the different experiments. This was particularly interesting

as previous experiments have focused upon the expression of its ligand CXCL12,

shown to be expressed on endosteal osteoblasts, vessels and reticular cells in vivo

(82, 196), and little emphasis has been placed on CXCR4 expression by osteoblasts. A

study by Ponomaryov et al (111) demonstrated low expression of CXCR4 by MBA-15

and BM stromal cells. In the studies in this chapter, there was no difference in the

expression of CXCR4 over-time when treated with OGM, but this may be due to the

heterogeneous nature of primary OLCs.

CXCL12 was also shown to be expressed by murine primary OLCs at each time point

of OGM treatment and was surprisingly high. Expression of CXCL12 by murine

osteoblasts is consistent with previous experiments which demonstrated the

expression of CXCL12 by the BM stroma and endosteal osteoblasts in vivo (82, 111).

There were no significant differences in the expression of CXCL12 over-time implying

that OGM treatment did not influence its expression. Interestingly, previous papers

have categorised CXCL12 expressing osteoblasts as immature osteoblastic cells

opposed to mature osteoblasts (111). This is supported by studies, which have

compared the levels of CXCL12 between MG-63 and SAOS-2 cells. Previous studies


reported that SAOS-2 cells were more mature with high mineralisation compared to

the MG63 cells which were described as immature and heterogeneous (196, 257).

When the CXCL12 levels were compared, the more mature SaOS-2 cells had lower

expression of CXCL12 compared to the immature MG-63 cells (111, 258). In addition,

Jung et al (196) demonstrated that human osteoblastic cells, MG63 cells, BM stromal

cells and MC3T3-E1 cells secreted SDF-1 in the early time points of differentiation

(detected by enzyme-linked immunosorbent assay in conditioned media) and

production of SDF-1 decreased over the time course, again highlighting that SDF-1 is

produced in the early stages of osteoblastic differentiation. With regards to the

experiments in this chapter, it may seem that the osteoblastic population used was

still quite heterogeneous, particularly as CXCL12 expression decreased after 7 days

and then reached maximal levels at 28 days. In addition, only mRNA was detected

therefore, protein production may have provided different results.

Notch-1 expression, was expressed by MC3T3-E1 osteoblastic cells (200), C3H10T1/2

and C2C12 cell lines and murine undifferentiated primary OLCs (216), therefore,

Notch-1 may be more prevalent in preosteoblastic cells rather than mature

osteoblasts. Notch-1 expression in the experiments in this chapter was detected in

osteoblasts at all time-points and no significant changes in Notch-1 expression over-

time were observed. Notch-1 expression did however, slightly decrease after 7 days

but this was variable. If the experiment was conducted for a longer period, a further

decrease may have been observed.

Jag-1 was expressed at similar levels to Notch-1 at all time-points. Jag-1 expression

has been determined in a variety of different osteoblastic cells including MC3T3-1 E1

(200), BM stromal cells in vitro (199) and osteoblastic cells in vivo (82, 259). Jag-1

expression levels by osteoblastic cells has not been associated with osteoblast

differentiation states. Although, stromal cells and MC3T3-E1 cells are usually classed

as immature osteoblasts or preosteoblasts, therefore, this correlates with the trend


observed whereby Jag-1 expression decreased from 7 days onwards in the

experiments in this chapter.

Tie-2 was expressed by the primary OLCs at each time-point. Its expression was

particularly low and did not change over-time. Tie-2 is typically a marker of ECs and

has not been a key focus in osteoblastic cells. However, low expression of Tie-2 was

determined in MC3T3-E1 cells, murine primary OLCs, ST-stromal cells, C2C12 and

C3H10T1/2 cells (202). Overall Tie-2 expression was higher in the C3H10T1/2 cells

(202), which may imply that Tie-2 may be more prevalent in early mesenchymal

progenitor cells, rather than immature or mature osteoblastic cells. However, this

was contradicted in studies using primary osteoblastic cells and C2C12 cells, which

seemed to demonstrate increased expression of Tie-2 following three days of culture

in OGM, therefore, Tie-2 is present at higher levels in more differentiated cells in this

instance (202). However, in the experiments in this chapter, no pattern of Tie-2

expression was observed across the time course of differentiation, but again this

could be due to the heterogeneity in the population.

Ang-1 was expressed over-time at similar levels to Notch-1, Jag-1 and N-cad. Jeong

et al (202) demonstrated an initial increase in Ang-1 expression within the first three

days of primary murine osteoblastic cell differentiation compared to day 0 and this

was then maintained at the same level for a further 25 days. In these experiments,

differences in Ang-1 expression may not have been determined, as 7 days was the

earliest time point used and any increase in Ang-1 expression may have already

occurred.

N-cad was expressed by the primary OLCs over-time. After 14 days of differentiation

N-cad expression seemed to decrease however, this did not reach significance. N-cad

was previously reported to be expressed by a variety of different MSC lineage cells


such as MC3T3-E1 cells, C3H10T1/2, ST-stromal cells, ATDC5 chondrogenic cells,

NIH3T3 fibroblast cells (218) and KS483 osteoblastic cells (206). However, the effect

of differentiation upon N-cad is currently unclear. Greenbaum et al (260)

demonstrated expression of N-cad by primary murine osteoblastic cells which

diminished following 15 days of differentiation whereas, Ferrari et al (219)

demonstrated an increase in N-cad expression as SaOS-2 cells were differentiated

over the course of 4 days. It is therefore, difficult to ascertain whether N-cad

expression would increase upon differentiation stimuli. In the experiments in this

chapter, no significant difference in N-cad expression was found over-time with OGM

treatment. However, after 14 days there was a decrease in N-cad expression and

again if the experiment was continued for a longer period, further declines may have

been observed, correlating with the experiments conducted by Greenbaum et al

(260).

Again, as these experiments lacked the suitable control of OLCs cultured in standard

medium as a comparator at each of the times points it is difficult to draw substantial

conclusions. For example, Col1α and Runx-2 levels did not significantly change over-

time however; the levels of these may be significantly increased or decreased in OLCs

cultured in OGM compared to OLCs cultured in control media, and similarly, the

levels of HSC niche molecules and ligands may also be have significantly different

between differentiating and non-differentiating OLCs. However, without the suitable

non-OGM controls the actual effect of OGM and therefore differentiation upon bone

marker or HSC niche molecule or ligand expression was unable to be suitably tested.

The adhesion of MCs to osteoblastic cells after treatment with OGM was then

determined. A higher number of 5TGM1-GFP cells adhered to the primary OLCs after

6 hours compared to 1 hour although, this did not reach significance. In addition,

significantly more MCs adhered to primary OLCs treated with OGM for 3 days

compared to 28 days after 6 hours. This was quite interesting as it implies that MCs,


like HSCs, seem to preferentially adhere to osteoblastic cells in the early stages of

differentiation, which is supported by Zhang et al (83) and Xie et al (84) when defining

osteoblastic cells in the HSC niche. These results may also imply that the factors

potentially responsible for this adhesion are not the HSC niche molecules or ligands

described in this thesis, as no differences in their expression was seen over-time.

However, as the day 3 time-point was not assessed by real-time PCR and as the

standard media controls were not used, this cannot be completely concluded.

There are several disadvantages related to subjectivity using this adhesion technique

that need to be noted. Firstly, washing the cells can severely affect the results, as

washing the cells with slightly different forces may either detach or retain adhesion.

In addition, results may be subjective as only 5 out of a 25 tile area were imaged,

which was particularly subjective as the majority of cells were found around the

edges of the wells compared to the centre, despite this, the same areas were imaged

in each well. Also at 28 days, the primary OLCs had begun to detach at the edges of

the wells, resulting in less area for the cells to adhere to, which may have also

influenced the results. Mineralisation may have also interfered with MC adhesion,

which was greatest at 21 days and again this could have provided less area for the

cells to adhere to. In addition, it may also have been desirable to examine the effect

of using PBS compared to media. PBS was previously used when investigating the

adhesion of HSCs and osteoblastic cells (261) for a short adhesion time of 15 min,

although 6 hours may be too long a time for MCs to survive in PBS as they have a

short life span when not in media.

Survey of the literature did not provide a standardised time for the adhesion assays.

Some experiments used 20 min (206), 2 hours, (262), 4 hours (263) or 6 hours (42).

Therefore, an early (1 hour) and late (6 hours) time point was chosen. However, in

retrospect, it may have been more desirable to look at the adhesion within the first

15 min, which could capture early simple associations and possibly even later than 6


hours such as 12 or 24 hours, which could potentially detect more complex

interactions involving gene expression in interacting cells. In addition, the number of

cells plated onto osteoblasts varied. Some studies used 1x105 (206, 261) and

5x106/ml (263) in a 96 well plate. However, in these experiments 5x104 cells were

used, as the 96 well plates did not have the capacity for greater numbers than this.

In conclusion, I have validated a model for the differentiation of murine primary OLCS

in vitro. Some of the characteristics were similar to what has been previously

reported using rat calvarial cells, BM stromal cells and osteoblastic cells however,

there were also some inconsistencies particularly with respect to the mineralisation

time points and peak expression of collagen. Treatment with OGM did not affect the

expression of HSC molecules or ligands but despite this, MCs adhered highly to

primary OLCS in the early time points of OGM treatment. Therefore, this may imply

that other molecules besides the HSC niche molecules and ligands investigated here

are important with regards to MC adhesion to osteoblasts, though as stated previous

this isn’t conclusive due to a lack of standard media controls.

Further research is now required to explore differences in the expression of other

adhesion molecules using this system and also to further optimise the adhesion

experiments. The assays used to determine the differentiation state of the BM

cultures were population based and it would be have been helpful to assess

differentiation in a per cell basis to determine the maturity of interacting cells. This

was however beyond the scope of the present study, but should be considered in

future studies.

Chapter VI: The effect of knocking-down HSC cell niche molecules in myeloma cells upon adhesion to osteoblast lineage cells in vitro and tumour burden in vivo Page 192

Chapter VI: The effect of knocking down HSC niche

molecules in myeloma cells upon adhesion to

osteoblast lineage cells in vitro and tumour burden in

vivo


6.1 Introduction

In the studies described in Chapter III-V, I determined the expression of several HSC

niche molecules by the 5T33MMvt and 5TGM1-GFP cells. These studies collectively

showed that MCs produced the HSC niche molecules CXCR4, Notch-1, Tie-2 and N-

cad and the ligands CXCL12 and Jag-1. Expression of these molecules by MCs has

previously been reported in the literature; however, the role of these molecules in a

myeloma osteoblastic niche is yet to be determined.

Previous preclinical studies have shown the importance of the CXCR4/CXCL12 axis for

its role in the mobilisation of CXCR4 positive HSCs (94, 112) and various cancer cells

including prostate (258, 264), breast (265, 266) and MCs (170, 194) towards a high

concentration gradient of CXCL12 in vitro and in vivo. In MM, in vitro transwell

experiments showed that 5T2MM and 5T33MM cells specifically migrated towards

fibroblast conditioned media containing CXCL12 and recombinant CXCL12. This was

significantly inhibited using the peptide antagonist 4F-benzoyl-TN14003 (170). These

observations were also supported by studies using human MC lines and primary

patient samples, where recombinant CXCL12 or conditioned media induced

chemotaxis (193, 194, 221) which was blocked by an anti-CXCR4 inhibitory antibody

or the CXCR4 antagonist AMD3100 (195).

The interactions of Notch-1 and Jag-1 have also been investigated in MM. Activation

of Notch-1 by Jag-1 and its effect upon cell proliferation is somewhat contradictory

in the literature. Jundt et al (198) described a 2-fold increase in MC proliferation

when stimulated with Jag-1, which was inhibited using the γ secretase inhibitor N-

[(3,5-Difluorophenyl)acetyl]-L-al-anyl-2-phenylglycine-1,1-dimethylethyl ester

(DAPT). Whereas, Nefedova et al (199), described a reduction in proliferation and an

increase in cells in the G0 phase of cycling when cells were co-cultured with Jag-1

expressing stromal cells. In addition, the interaction of Jag-1 positive stromal cells


with MCs propagated chemoresistance when melphalan was administered (199).In a

separate study, administration of a γ secretase inhibitor in combination with

doxorubicin, also resulted in MC chemosensitisation and increased the percentage

of apoptotic cells in vitro and in vivo (267).

N-cad expression and function has been implicated in several cancers including

breast, prostate and melanoma, particularly in relation to metastasis. Hazan et al

(268) demonstrated that N-cad expression in MCF-7 breast cancer cells induced both

migration and invasion in vitro and an increase in metastasis upon injection in vivo.

This was also supported by Nieman et al (204), who demonstrated that N-cad positive

breast cancer cell lines had higher invasion and motility. In prostate cancer, 93% of

patients with established metastasis had tumours which were positive for N-cad,

whereas only 56% of patients who had no known metastasis had N-cad positive

tumours (228). In addition, when N-cad was knocked-down in melanoma cells, this

resulted in a reduction in transwell migration in vitro (269). Furthermore, when N-

cad was blocked using an inhibitory antibody, this resulted in a reduction in

melanoma cell adhesion to fibroblasts and endothelial cells, a reduction in migration

towards fibroblasts in vitro and an increase in melanoma cell death (270).

N-cad has also been implicated in MC adhesion to osteoblastic cells and in the early

stages of MC homing and survival in the BM. NCIH929 MCs adhered to recombinant

N-cad and N-cad expressing C3H10T1/2 osteoblastic cells in vitro and this interaction

was subsequently blocked using an N-cad inhibitory antibody (206). The KD of N-cad

in NCIH929 cells also resulted in a reduction of MC homing or survival in the BM and

an increase in the number of cells in the circulation when injected into mice (206). In

addition, Sadler et al (207) demonstrated that inhibition of MC adhesion to stromal

cells in culture using an anti-N-cad antibody resulted in increased proliferation of

MCs, therefore, N-cad signalling may be important for cell proliferation and control

of cell cycle.


As stated previously, Tie-2 and Ang-1 have not been highly studied in MM. They have

activities typically associated with blood vessel development. Tie-2 is highly

expressed in blood vessels and is required for vessel stability and integrity, which is

stimulated by Ang-1 and inhibited by Ang-2. However, vessel stability is an important

mechanism for the establishment of solid tumours and Vacca et al (271) reported an

increase in micro-vessel area in myeloma patients. Therefore, Tie-2/Ang-1

interactions may be important in myeloma establishment in the BM.

Successful interventions of all the molecule interactions discussed above in disease

have now also been translated into clinical trials for various types of cancer, discussed

in detail in the general discussion, Chapter VII.

In this Chapter, I have used a shRNA gene KD strategy to supress the expression of

genes for CXCR4 and N-cad in 5TGM1-GFP cells. I then conducted functional studies

to determine the effect of the KD on the adhesion of the MCs to the differentiated

primary OLCs (optimised in Chapter V) and also the effect upon tumour growth in

vivo in a pilot study. These experiments were conducted to determine the role of

these molecules in a MC niche and to identify whether these molecules are important

targets for future intervention studies. The 5TGM1-GFP model was used, as it is

intrinsically less variable in terms of tumour burden and tumour take compared to

the 5T33-GFP model. The 5T33-WT model was also used in this chapter as a control

to test the transduction of GFP using the lentiviral system.


6.2.1 Hypothesis and aims

The aim of these studies was to test the hypothesis that “The knock-down (KD) of

HSC niche molecules in MM cells, results in reduced myeloma cell (MC) adhesion to


primary osteoblast lineage cells (OLCs) in vitro and subsequent reduction in tumour

burden in vivo”.

6.2.2 Objectives


1. To KD HSC niche molecules in the 5TGM1-GFP cells using short hair-pin RNA

(shRNA) and to transduce green fluorescent protein (GFP) into 5T33-WT cells.

Gene KD will be validated using real-time PCR and Western blotting (WB), and

GFP transduction will be confirmed using flow cytometry.

2. To determine the effect of knocking down HSC niche molecules on MC adhesion

to primary OLCs in vitro.

3. To determine the effect of knocking down HSC niche molecules upon tumour

burden in vivo.

6.3 Methods

6.3.1 Short hair-pin RNA knock-down strategy

6.3.1.1 Puromycin killing curves

5TGM1-GFP cells and 5T33-WT cells were seeded in triplicate at a density of 1x105

cells into a 12 well plate in 2 ml complete RPMI media containing either 0 (an equal

volume of PBS), 0.5, 1.0, 2.5, 5.0, 7.5 and 10 µg/ ml puromycin and after 2 and 4 days

the cell viability was determined using TO-PRO-3 staining followed by flow cytometry

using a FACS Calibur and Cell Quest software. The lowest dose required for 100% cell

death over a 4-day period was then established by analysing the number of viable

cells remaining at the different concentrations of puromycin at 2 and 4-day time

points. Results were based on three independent experiments (N=3). This is

described in more detail in Chapter II, Section 2.6.2.2.


6.3.1.2 Polybrene optimisation

5TGM1-GFP and 5T33-WT cells were seeded in triplicate into a 48 well plate at a

density of 5x104 in 500 µl complete RPMI media containing 0 (PBS of the same

volume), 2.5, 5.0, 7.5 and 10.0 µg/ml polybrene and incubated for 24 hours. After 24

hours, the number of cells in each well were counted using a haemocytometer to

establish whether high concentrations of polybrene were toxic. Results were

established from three independent experiments (N=3). This was described in more

detail in Chapter II, Section 2.6.3.1.

6.3.1.3 Lentiviral green fluorescent protein transduction

The 5T33-WT cells were seeded at a density of 2x104 cells in 100 µl complete RPMI

media containing 5.0 µg polybrene (determined from the optimisation experiments)

in a 96 well plate and the lentiviral particles for GFP and scrambled controls at a

multiplicity of infection (MOI) of 5 (CTRL 5), 15 (CTRL 15) and 25 (CTRL 25) were added

and then incubated for 20 hours. Green fluorescent protein expression was

determined in puromycin resistant colonies (treated with 7.5 µg puromycin) using

flow cytometry. The stability of the GFP transduction was also established in cells in

the absence of puromycin free media, four weeks after transduction. This was

described in further detail in Chapter II, Section 2.6.4.1.

6.3.1.4 Lentiviral short hair-pin RNA transduction for CXCR4 and N-cad

The 5TGM1 cells were seeded at a density of 2x104 cells in 100 µl complete RPMI

media containing 5.0 µg polybrene in a 96 well plate and the lentiviral particles

containing shRNA for N-cad and CXCR4 at MOIs of 5 (KD 5), 15 (KD 15) and 25 (KD 25)

or scrambled controls at the same MOIs were added and then incubated for 20 hours.

Knock-down was detected in puromycin resistant colonies (treated with 5.0 µg/ ml

puromycin) using real-time PCR and WB. The cells were then cultured in the absence

of puromycin for four weeks to detect the stability of the KD, which was analysed


using real-time PCR and WB. This was described in further detail in Chapter II, Section

2.6.4.1.

6.3.2 Assessment of gene knock-down

6.3.2.1 Real-time PCR

For PCR analysis of the KD, 1x106 CXCR4 and N-cad KD 15 and 25 and CTRL 15 and 25

treated cells were taken in triplicate for RNA extraction. RNA was extracted, cDNA

was synthesised and real-time PCR was conducted as described in Chapter II, Sections

2.3.1.1, 2.3.2.1 and 2.3.5.2. Real-time PCR was analysed using the comparative ΔΔCt

method by normalising to the HK gene β2M and the percentage of KD was calculated

as described in Chapter II, Section 2.6.4. Results were established from three


6.3.2.2 Western blotting

Western blotting was used to confirm the KD of protein in the 5TGM1-GFP cells. In

each sample, 1x107 CTRL 15 and 25 and N-cad KD 15 and 25 cells were taken in

triplicate and lysed using NP-40 buffer. The protein was quantified using a

bicinchoninic acid (BCA) assay and run in a 7% resolving gel using 30% acrylamide by

polyacrylamide gel electrophoresis (SDS-PAGE). The gel was transferred onto a

nitrocellulose membrane and N-cad protein was determined using sandwich

“immuno-blotting”. Results were visualised by chemoluminescence and quantified

using ImageJ software to calculate band density. Knock-down was quantified by

normalising to the HK protein β-actin and a percentage of protein KD was calculated.

Results were established from three independent experiments (N=3). This is

described in more details in Chapter II, Section 2.4.4.


6.3.3 Assessment of the in vitro effects of HSC niche molecule gene knock-down in

transduced myeloma cells

6.3.3.1 Growth curves

To assess the effect of the HSC niche molecule KD upon cell proliferation in vitro,

growth curve experiments were conducted, where 5x103 cells (5TGM1-GFP, CTRL 15

and 25 and KD 15 and 25) were seeded into a 96 well plate in 200 µl complete RPMI

media. Cell counts were recorded every 2 days for 12 days. Cell proliferation was

presented in a graphical format to visualise each phase of growth and the doubling-

time in the exponential phase was calculated using the formula shown in Chapter II

section 2.4.5.1. Results were established from three independent experiments (N=3).

6.3.3.2 Adhesion experiments

To assess the effect of the HSC niche molecule KD upon MC adhesion to primary

OLCs, the same method was used as developed previously in Chapter V. In brief,

5x104 cells (CTRL 15 and 25 and N-cad KD 15 and 25) were seeded in 200 µl complete

RPMI media onto primary OLCs differentiated in 10 mM BGP and 50 µg/ml asc for 3,

14 and 28 days in a 96 well plate. After 1 and 6 hours, plates were washed and the

remaining cells that adhered to the primary OLCs were visualised using a Leica

AF6000 fluorescent microscope and LAS LF software. This is described in more detail

in Chapter II, Section 2.5.3.2.

6.3.4 Assessment of the in vivo effects of HSC niche molecule gene knock-down in

transduced myeloma cells

A pilot in vivo experiment using the N-cad KD cells was conducted to firstly, identify

any differences in tumour burden between the transduced cells and standard

5TGM1-GFP cells (as it was suspected that prolonged in vitro culture may effect

tumourgenicity) and secondly, to determine any differences in tumour burden

between mice injected with CTRL 25 and N-cad KD 25 cells. Two million 5TGM1-GFP


cells (N=4), CTRL 25 (N=6) and KD 25 (N=6) cells were injected into 11 week old

female C57BL/KaLwRij mice. When the first mouse became ill (at 25 days post

injection), all mice were sacrificed and the hind limbs were dissected. The MM

models are described in more detail in Chapter II, Section 2.2.2.1.

6.3.4.1 Tumour burden analysis

Both femurs from each mouse were flushed using PBS and analysed separately by

flow cytometry to determine the percentage of GFP positive tumour cells in each

femur as described in Chapter II, Sections 2.2.1.4 and 2.4.1.2. In addition, the left

tibiae were fixed in 4% PFA, decalcified in PFA/EDTA and sections were cut and

stained by CD138 IHC as described in Chapter II, Section 2.4.3.3. The stained slides

were scanned using an Aperio Scan Scope slide scanner and tumour area and colony

number were analysed manually using Image Scope software.

6.3.4.2 Bone disease analysis

The right tibia from each mouse was fixed in 10% formalin and scanned by µCT using

a SkyScan 1172 at 50 Kv, 200 µA, an aluminium filter of 0.5 mm and pixel size of 4.3

µm2. Images were reconstructed using N-Recon Software. Trabecular bone 0.2 mm

below the growth plate was analysed using Ct-an and Batman software to provide

data for bone parameters including BV/TV, tb.n and tb.th. This is described in greater

detail in Chapter II, Section 2.7.1.2. Lesion number and area were also calculated

using Ct-an software by utilising reconstructed µCT images of each tibiae. Following

this, the images were resized; the cortical bone region of interest was selected for

lesion analysis using ImageJ software. This is described in greater detail in Chapter II,

Section 2.7.1.2.


6.3.5 Statistics

Non-parametric statistical tests were used where relevant. A Friedman test followed

by Dunn’s post test was used to determine differences in cell death between MCs

treated with different concentrations of puromycin and polybrene and was also used

to determine differences in GFP expression in MCs cultured in the absence of

puromycin over-time. A Kruskal-Wallis test followed by a Dunn’s post test was used

to determine differences in cell proliferation between CTRL and KD cells over-time.

A Mann-Whitney U test was used to compare differences in CTRL and KD MC

adhesion to primary OLCs cultured in OGM for 3, 14 and 28 days and a Friedman test

was used to determine differences in MC adhesion to primary OLCS cultured in OGM

for 3, 14 and 28 days. In addition, a Mann-Whitney U test was used to compare

tumour burden and bone disease between CTRL 25 and N-cad KD injected mice in

vivo. A Fisher’s exact test was also used to compare the tumour take.

Where relevant, error bars were displayed on all graphs using the SEM unless stated

otherwise in the legend.

6.4 Results

6.4.1 Killing curves to determine the optimum concentration of puromycin using

flow cytometry

Puromycin killing curves were conducted to determine the optimum concentration

of puromycin to select for resistant colonies following lentiviral transduction for the

KD of CXCR4 and N-cad by shRNA. Cell death was analysed by TO-PRO-3 staining using

flow cytometry where non-viable cells were positive.

Figure 6.1 demonstrates the percentage of 5TGM1-GFP and 5T33-WT cell death after

2 and 4 days using increasing doses of puromycin. As the concentration of puromycin

increased, the percentage of TO-PRO-3 positive 5TGM1-GFP and 5T33-WT cells also

increased. After 2 days, the number of TO-PRO-3 positive 5TGM1-GFP cells increased

between 0.0-5.0 µg/ml puromycin after which cell death plateaued at the remaining

concentrations. Cell death was significant at 5.0, 7.5 and 10.0 µg/ml puromycin

compared to the control (P<0.05). After 4 days again there was an increase in TO-

PRO-3 positive cells, but the cell death plateaued earlier at 2.5 µg/ml, after which

there was a 2% increase in cell death between 2.5 and 5.0 µg/ ml puromycin. The

mean percentage of cell death of 97% then remained constant for the remainder of

the concentrations. Statistically, cell death was significant at 5.0 µg/ ml puromycin

compared to the control (P<0.05).

After 2 days, the 5T33-WT cells demonstrated an increase in TO-PRO-3 positive cells

from 0.0-5.0 µg/ml. This was then followed by a 5% increase between 5.0 µg/ml and

7.5 µg/ml and a 2% increase between 7.5 µg/ml and 10.0 µg/ml. Statistically, there

was a significant difference in the number of dead cells at 7.5 µg/ml (P<0.05) and

10.0 µg/ml (P<0.01) compared to the control, and also between 0.5 µg/ml and 10.0

µg/ml (P<0.05). After 4 days, the number of TO-PRO-3 positive cells increased

between 0.0-5.0 µg/ml followed by a 2% increase between 5.0 µg/ml and 7.5 µg/ml

and no increase between 7.5 µg/ ml and 10.0 µg/ml puromycin. Statistically, there

was a significant difference in cell death between 7.5 µg/ml (P<0.01) and 10.0 µg/ml

(P<0.05) puromycin compared to the controls.

Figure 6.1: Increasing doses of puromycin resulted in an increase in 5TGM1-GFP and 5T33-WT cell death. 5TGM1-GFP and 5T33-WT cells were seeded at 1x105 cells per 12 well plate in 2 ml of complete RPMI media and incubated over-night. The following day the cells were centrifuged at 300 x g for 5 min and resuspended in 2 ml of complete RPMI media containing 0.0, 0.5, 1.0, 2.5, 5.0, 7.5 and 10.0 µg/ml puromycin. After 2 and 4 days the number of dead cells were detected by TO-PRO-3 staining and flow cytometry using a FACS Calibur and Cell Quest software. a. 5TGM1-GFP cell death after 2 and 4 days using increasing doses of puromycin. b. 5T33-WT cell death after 2 and 4 days using increasing doses of puromycin. Data was analysed statistically using a Friedman test followed by Dunns post test (P<0.05) to determine differences in TO-PRO-3 staining using the different doses of puromycin. Results were determined from four independent experiments (N=4).

From the data, it was also clear that a higher number of cells were TO-PRO-3 positive

after 4 days compared to 2 days using each concentration of puromycin and cell type.

a

b

In addition, the 5T33-WT cells seemed to be more resistant to the puromycin

compared to 5TGM1-GFP cells as higher concentrations of puromycin were required

to kill the same number of cells. However, due to a lack of non-parametric two-way

paired test, neither of these findings were analysed statistically.

From these experiments, I was able to determine the suitable dose of puromycin for

use in future experiments, required to select for resistant colonies in the KD

experiments. In future experiments, I used a dose of 5.0 µg/ml of puromycin to select

resistant colonies using the 5TGM1-GFP cells and 7.5 µg/ml of puromycin to select

for resistant 5T33-WT cells. Both of these concentrations produced significantly

greater cell death compared to the controls. In addition, the manufacturer

recommended using a concentration which kills 100% of cells within four days and

using these concentrations in my experiments a plateau in cell death was

demonstrated with no increase in cell death after use at these concentrations.

6.4.2 The effect of polybrene upon cell viability after 24 hours of treatment

The effect of polybrene upon cell viability was tested after 24 hours of treatment at

several different dose from 0.0 (PBS vehicle control), 2.5, 5.0, 7.5 and 10.0 µg/ml

polybrene, by cell counting using a haemocytometer. Cell counts were used opposed

to TO-PRO-3 staining, as polybrene acts by increasing pore sizes and upon using TO-

PRO-3 and flow cytometry it resulted in false positives for dead cells.

Figure 6.2 shows the effect of polybrene upon cell number. Statistically, polybrene

had no significant effect upon cell viability using any concentrations between 0.0 and

7.5 µg/ml in the 5TGM1-GFP cells, however, 10.0 µg/ml of polybrene resulted in a

significant decrease in cell number (P<0.05) after 24 hours. Polybrene however, had

no effect upon 5T33-WT cell number after 24 hours at any dose (0.0-10.0 µg/ml). As

polybrene had no significant effect on proliferation at any of the doses using the

5T33-WT cells and no affect upon proliferation up to 7.5 µg/ml using the 5TGM1-GFP

cells, the recommended dose by the manufacturer of 5.0 µg/ml was used for future

experiments.

Figure 6.2: Low doses of polybrene had no effect upon 5TGM1-GFP or 5T33-WT cell proliferation after 24 hours. 5TGM1-GFP and 5T33-WT cells were seeded at 5x104 cells per 48 well plate in 500 µl complete RPMI media containing 0.0 (PBS vehicle control), 2.5, 5.0, 7.5 and 10.0 µg/ml polybrene. After 24 hours the number of cells in each well, at each concentration (x-axis), were counted using a haemocytometer (y-axis). Data was analysed statistically using a Friedman test followed by Dunn’s post test (P<0.05) to determine differences in cell number following polybrene treatment. Results were determined from four independent experiments (N=4).

6.4.3 Green fluorescent protein transduction in the 5T33-WT cells as a proof of

principal

A lentiviral GFP transduction system was used as a proof of principal to test whether

lentiviral transduction using these particular reagents was possible in the MCs. The

5T33-WT cells were chosen, as we did not possess the non-GFP transduced 5TGM1

variety.

Lentiviral particles containing the GFP sequence and puromycin resistance vector

were transduced into the polybrene (5.0 µg/ml) treated 5T33-WT cells at MOIs of 5,

15 and 25. This was followed by puromycin selection at 7.5 µg/ml, optimised as


shown in Chapter VI, Section 6.3.1, Figure 6.1. The cells treated with an MOI of 5 did

not survive following puromycin selection whereas, those treated with an MOI of 15

and 25 were viable, indicating that transduction of the puromycin resistance vector

was successful at the higher MOIs in the 5T33-WT cells.

Flow cytometry was used to detect the success of the GFP transduction in the 5T33-

WT cells following puromycin selection (0 days). To determine the stability of the GFP

transduction, cells were cultured in puromycin free media for 7, 14, 21 and 28 days,

followed by flow cytometry. As shown in Figure 6.3, the 5T33-WT cells, transduced

using both MOIs, expressed GFP after puromycin selection and also in puromycin free

media over-time. Statistically, there were no significant differences in the expression

of GFP over-time using MOIs of 15 and 25.

Figure 6.3: Green fluorescent protein was susccessfully transduced into 5T33-WT cells using lentiviral particles and analysed by flow cytometry. 5T33-WT cells were seeded at 2x104 cells per 96 well plate in 100 µl complete RPMI media containing 5.0 µg/ml polybrene and 15 or 25 MOI of lentiviral GFP particles. After 20 hours the cells were washed twice in complete RPMI media and resuspended in complete RPMI media for 72 hours, after which they were treated with 7.5 µg/ml puromycin. Once selected colonies appeared, 1x106 cells were taken for flow cytometry (day 0) after which the cells were washed twice in RPMI media and cultured in complete media in the absence of puromycin for 7, 14, 21 and 28 days and at each time point the number of cells positive for GFP per 1x106 sample was assessed (y-axis) using a FACS Calibur and Cell Quest software. (x-axis). Data was analysed using a Friedman test followed by Dunn’s post test (P<0.05) to compare differences in GFP expression by the MOI 15 and 25 treated cells over-time. This experiment was conducted only once but set-up in triplicate.

6.4.4 Real-time PCR to establish the success of knocking-down CXCR4 and N-cad

using the lentiviral short hair-pin RNA transduction system

The lentiviral shRNA system was used to KD the genes for CXCR4 and N-cad. Optimum

conditions established using the 5T33-WT cells for GFP transduction were used. The

5TGM1-GFP cells were treated with polybrene and the lentiviral particles at MOIs 15

and 25. This was followed by puromycin selection at 5.0 µg/ml.

The KD of CXCR4 and N-cad was determined using real-time PCR. Figure 6.4

demonstrates the effect of lentiviral shRNA transduction upon gene expression of

CXCR4 (Figure 6.4a) and N-cad (Figure 6.4b) at MOIs of 15 and 25 after puromycin


selection. When the KD of CXCR4 was assessed, there was only a 24% reduction in

CXCR4 expression at an MOI of 15 and no difference using the MOI of 25 compared

to the control. At this stage, and after a number of attempts, it was clear that the KD

of CXCR4 in the 5TGM1-GFP cells was unsuccessful, as assessed by real-time PCR.

Therefore, the CXCR4 shRNA transduced cells were not used in future experiments.

In contrast to CXCR4, when the KD of N-cad was assessed, there was a reduction in

N-cad gene expression of 70% using the MOI of 15 and 71% using the MOI of 25,

compared to their corresponding controls. According to the manufacturer’s

instructions, a successful KD is achieved if the KD is greater than 60%, therefore, KD

of N-cad was successful at both MOIs and the N-cad shRNA transduced 5TGM1-GFP

cells were used in future experiments.


Figure 6.4: N-cad was successfully knocked-down in the 5TGM1-GFP cells using short hair-pin RNA, analysed by real-time PCR. 5TGM1-GFP cells were seeded at 2x104 cells per 96 well plate in 100 µl complete RPMI media containing 5.0 µg/ml polybrene and 15 or 25 MOI of CTRL, CXCR4 or N-cad lentiviral particles. After 20 hours, the cells were washed twice in complete RPMI media and resuspended in complete RPMI media for 72 hours, after which they were treated with 5.0 µg/ml puromycin. Once selected colonies appeared, 1x106 cells were taken for RNA extraction, cDNA synthesis and real-time PCR for N-cad and CXCR4. a. The KD of CXCR4 in the 5TGM1-GFP cells using CTRL or CXCR4 KD shRNA lentiviral particles at a MOI of 15 or 25. b. The KD of N-cad in the 5TGM1-GFP cells using CTRL or N-cad shRNA lentiviral particles at a MOI of 15 or 25. Data was analysed using the comparative ΔΔCt method relative to the HK gene β2M. Gene expression is shown using the fold change (y axis) and the percentage of gene expression is above the columns. Results were determined from three independent experiments (N=3) and presented using the SD.

a

b


6.4.5 Real-time PCR to establish the stability of N-cad knock-down in cells cultured

in puromycin free media

Experiments were conducted to assess the stability of the N-cad KD in cells no longer

cultured under puromycin selection. The rationale for this was to ensure that cells

remained knocked-down for N-cad in the absence of puromycin because once

injected in vivo they would no longer be under selection conditions.

Cells were cultured in vitro for 28 days without puromycin selection, followed by PCR

analysis. This time-point was chosen due to the length of the 5TGM1-GFP in vivo

model. Figure 6.5 demonstrates the effect of culturing the 5TGM1-GFP N-cad KD cells

in the absence of puromycin after 28 days, upon N-cad gene KD. N-cad expression

was reduced by 54% using a MOI of 15 and 75% using a MOI of 25. Therefore, after

MC growth in complete media without puromycin, cells treated with a MOI of 15

were no longer suitably knocked-down as the reduction was no longer greater than

60% whereas, cells treated with an MOI of 25 remained successfully knocked-down

with a reduction greater than 60%.


Figure 6.5: N-cad gene expression was stably knocked-down in the 5TGM1-GFP cells cultured in the absence of puromycin for 28 days, analysed by real-time PCR. 5TGM1-GFP CTRL 15 and 25 and KD 15 and 25 cells were cultured in puromycin free complete RPMI media for 28 days after which 1x106 cells were taken for RNA extraction, cDNA synthesis and real-time PCR for N-cad. Data was analysed using the comparative ΔΔCt method relative to the HK gene β2M. Gene expression is shown using the fold change (y axis) and the percentage of gene expression is above the columns. Results were determined from three independent experiments (N=3) and presented using the SD.

6.4.6 Western blotting to establish the success of knocking-down N-cad using the

lentiviral short hair-pin RNA transduction system

To further quantify the KD of N-cad in the 5TGM1-GFP KD cells, WB was used. Figure

6.6 shows the effect of shRNA treatment in the 5TGM1-GFP cells upon N-cad protein

production at MOIs of 15 and 25 compared to the corresponding controls. Western

blotting clearly shows a decrease in protein for N-cad at an MOI of 15 and 25, based

upon band density (Figure 6.6a). When band density was analysed (Figure 6.6b) the

N-cad KD cells at a MOI of 15 demonstrated a KD in protein of 68% whereas, the KD

25 cells demonstrated a KD of 65%. As with the real-time PCR data, N-cad KD was

successful as KD was greater than 60%.


Figure 6.6: N-cad protein was successfully knocked-down in the 5TGM1-GFP cells using short- hair-pin RNA, analysed by Western blotting. N-cad was knocked-down in the 5TGM1-GFP cells using N-cad shRNA at a MOI of 15 and 25 followed by selection using 5.0 µg/ml puromycin. Once selected colonies appeared, 1x107 cells were lysed in NP-40 buffer, 20 µg protein was loaded into a 7% acrylamide gel and run by SDS-PAGE. The gel was transferred onto a

nitrocellulose membrane and immuno-blotted for N-cad and the HK β-actin. a. Representative image of a WB showing the protein for N-cad in the CTRL 15 and 25 lysates compared to KD 15 and 25. The size of the proteins are shown on the left. b. Western blotting analysis of N-cad protein density, calculated using ImageJ software. Data was analysed by normalising the protein band density of N-cad to the HK β-actin (y-axis) followed by comparative WB analysis to produce a percentage of N-cad protein density in 5TGM1-GFP cells

using the different MOIs (above graphs). Results were determined from three independent experiments (N=3).

6.4.7 Western blotting to establish the stability of N-cad knock-down in cells cultured in puromycin free media To determine the stability of N-cad protein KD in the 5TGM1-GFP cells, the cells were

cultured for 28 days in the absence of puromycin, followed by WB analysis. Figure

6.7a demonstrates the WB of the CTRL and KD 5TGM1-GFP cells at MOIs of 15 and

25. Visually from the WB there was a reduction in N-cad protein in the cells at both

the MOI of 15 and 25. Figure 6.7b demonstrates the quantification of N-cad protein

KD in the 5TGM1-GFP cells at MOIs of 15 and 25. N-cad protein was reduced by 49%

using the MOI of 15 and 60% using the MOI of 25. Therefore, as with the real-time

a

b


PCR analysis, the cells treated with a MOI of 15 and cultured in the absence of

puromycin for 28 days, no longer have suitable protein KD for N-cad (reduction was

less than 60%). Whereas, those treated with an MOI of 25 were still suitably knocked-

down, with 60% reduction in N-cad protein after 28 days.

Figure 6.7: N-cad protein was stably knocked-down in the 5TGM1-GFP cells cultured in the absence of puromycin, analysed by Western blotting. CTRL 15 and 25 and KD 15 and 25 cells were cultured in puromycin free complete RPMI media for 28 days. Ten million cells were then lysed in NP-40 buffer, 20 µg protein was loaded into a 7% acrylamide gel and run by SDS-PAGE. The gel was transferred onto a nitrocellulose membrane and immuno-blotted for N-cad and the HK β-actin. a. A representative image of a WB showing the protein for N-cad in the CTRL 15 and 25 lysates compared to KD 15 and 25. The size of the proteins are shown on the left. b. Western blot analysis of N-cad protein band density, calculated using ImageJ software. Data was analysed by normalising the protein density of N-cad to the HK β-actin (y-axis) followed by comparative WB analysis to produce a percentage of N-cad protein density in 5TGM1-

GFP cells using the different MOIs (above graphs). Results were determined from three independent experiments (N=3).

6.4.8 The effect of N-cad knock-down upon cell proliferation

In vitro experiments were conducted to determine any differences in cell

proliferation due to the KD of N-cad in the 5TGM1-GFP KD cells transduced using an

MOI of 15 and 25. Figure 6.8a demonstrates the growth curves produced over a 12-

b

a


day period using the un-transduced 5TGM1-GFP cells and CTRL and N-cad KD cells at

MOIs of 15 and 25, displayed using the cell number recorded at each time point. The

growth curves displayed different stages of proliferation, whereby there was an

initial lag phase between days 0-4, a short exponential phase between 4-8 days,

followed by a death phase from 8-12 days. A plateau phase was not observed within

these cells types, potentially due to how quickly the cells proliferate, which may

result in quicker depletions in media. The doubling time was then calculated from

the cells numbers in the exponential phase, as shown in Figure 6.8b. Statistically,

there were no significant differences in the doubling times between the standard

5TGM1-GFP cells, transduced control cells and N-cad KD cells.


Figure 6.8: N-cad knock-down in the 5TGM1-GFP cells had no effect upon cell proliferation. 5TGM1-GFP, CTRL 15 and 25 and N-cad KD 15 and 25 cells were seeded at 5x103 cells into a 96 well plate in 200 µl complete RPMI media without puromycin for the standard 5TGM1-GFP cells and with 5.0 µg/ml of puromycin for the CTRL and KD cells. Cells were cultured over a 12 day period and counted every 2 days using a haemocytometer. a. Growth curves using cell number (y-axis) of the 5TGM1-GFP, CTRL 15 and 25 and N-cad KD 15 and 25 at each time point (x-axis). The different phases of growth are separated (black-dashed lines) and labelled above that phase b. The doubling time in the exponential phase of growth, calculated using a standard formula. Data was analysed statistically using a Kruskal-Wallis test followed by Dunn’s post test to compare the doubling times between the cell types. Results were determined from four independent experiments (N=4).

6.4.9 The effect of N-cad knock-down upon adhesion to primary osteoblast lineage

cells in vitro

To test the effect of knocking-down N-cad in the 5TGM1-GFP cells upon adhesion to

primary OLCs cells in vitro, using the same methodology developed in Chapter V,

CTRL 15, CTRL 25, KD 15 and KD 25 cells were seeded onto primary OLCs which had

been cultured in OGM for 3, 14 and 28 days.

a

b

The effect of knocking-down N-cad in the 5TGM1-GFP cells upon adhesion to primary

OLCs in vitro was compared between the CTRL 15 and KD 15 cells and the CTRL 25

and KD 25 cells after 1 hour and 6 hours, as shown in representative fluorescent

images in Figure 6.9 and quantitative analysis in Figure 6.10. There were no

significant differences between the adhesion of KD 15 cells to OLCs at all time-points

compared to CTRL 15 cells. There was however, a consistent but not significant

reduction in the adhesion of KD 25 cells to OLCs at all time-points compared to the

CTRL 25 cells. In addition, there was also a trend for less adhesion of CTRL 25 cells to

OLCs across the time course with <50% adhesion on day 28 compared to day 3 but

again these did not reach significance.


Figure 6.9: Knock-down of N-cad in the 5TGM1-GFP cells had no significant effect upon myeloma cell adhesion to primary osteoblast lineage cells in vitro. CTRL 15 and 25 and KD 15 and 25 cells were seeded (5x104 cells per well) on to primary OLCs cultured in OGM containing 10 mM BGP and 50 µg/ml asc for 3, 14 and 28 days. Adhesion after 1 and 6 hours was imaged using a Leica AF6000LX microscope at a 10X magnification using LAS LF software. a. i-iv. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 3 day differentiated OCLs after 1 hour. v-viii. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 3 day differentiated OCLs after 6 hours. b. i-iv. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 14 day differentiated OCLs after 1 hour. v-viii. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 14 day differentiated OCLs after 6 hours. c. i-iv. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 28 day differentiated OCLs after 1 hour. v-viii. Representative images of CTRL 15 and 25 and KD 15 and 25 adhering to 28 day differentiated OCLs after 6 hours. Scale bars represent 100 µM

Figure 6.10: Knock-down of N-cad in the 5TGM1-GFP cells had no significant effect upon MC adhesion to primary osteoblast lineage cells in vitro. Fifty thousand CTRL 15 and 25 and KD 15 and 25 cells were seeded on to primary OLCs cultured in OGM containing 10 mM BGP and 50 µg/ ml asc for 3, 14 and 28 days. Adhesion after 1 and 6 hours was imaged using a Leica AF6000LX microscope at a 10X magnification using LAS LF software and analysed using Volocity software to determine the number of MCs adhering to primary OLCs per tile. i-ii. A comparison between the number of CTRL 15 and KD 15 cells adhering to primary OLCs cultured in OGM for 3, 14 and 28 days after 1 hour (i) and 6 hours (ii). iii-iv. A comparison between the number of CTRL 25 and KD 25 cells adhering to primary OLCs cultured in OGM for 3, 14 and 28 days after 1 hour (iii) and 6 hours (iv). Data was analysed statistically using a Mann Whitney U test (P<0.05) to determine any differences in the number of CTRL and KD cells adhering to primary OLCs cultured in OGM for at 3, 14 and 28 days after 1 and 6 hours. A Friedman test followed by Dunn’s post test was used to determine any differences in MC adhesion to primary OLCs at 3, 14 and 28 days cultured in OGM (P<0.05). Results were determined from four independent experiments (N=4).

6.4.10 Flow cytometry to determine the effect of knocking-down N-cad in the

5TGM1-GFP cells upon myeloma cell growth in vivo

An initial pilot study was conducted to firstly, determine whether the transduced

CTRL 25 and KD 25 cells would home and grow in the BM of injected mice, as

extensive time in culture could reduce tumourgenicity in these cells. Secondly, it was

conducted to determine whether there was any difference in tumour load between

the CTRL and KD groups.

ii i

iv iii

The standard 5TGM1-GFP cells, CTRL 25 or KD 25 cells were injected into

C57BL/KaLwRij mice and all groups were sacrificed when one of the mice became ill,

which was at 25 days P-I in the 5TGM1-GFP injected group. The left and right femora

were dissected from each mouse and the BM was isolated to detect GFP positive

tumour cells by flow cytometry. Figure 6.11 shows representative flow cytometry

profiles of the BM from mice injected with standard 5TGM1-GFP cells, CTRL 25 and

KD 25 cells.

The GFP positive tumour cell number was assessed by flow cytometry, as shown in

Figure 6.11b. The data demonstrated a higher number of GFP positive tumour cells

in the BM of the 5TGM1-GFP injected animals (32.90% ± 3.7, data not shown)

compared to the CTRL 25 injected animals (0.6% ± 0.59) and GFP positive KD 25 cells

were not detected in the animals injected with KD 25 cells (0.01% ± 0.00) using flow

cytometry. Interestingly, there was a significant reduction in KD 25 cells present in

the BM of mice compared to CTRL 25 cell injected animals (P<0.05).

Tumour take was assessed using a Fisher’s exact test, to compare the number of mice

and number of bones (as two femurs were used) with positive tumour burden

between the groups. Bones with more than 0.05% GFP positive cells were deemed

positive for tumour burden. As shown in Figure 6.11c, each mouse in the 5TGM1-GFP

group was positive for GFP (100%), in the CTRL 25 injected group 4 out of 6 mice had

GFP positive tumour burden (67%) and no mice in the KD 25 demonstrated tumour

burden greater than 0.05% (0%). When compared using the Fisher’s exact test there

was a significant difference in the number of mice with tumour in the 5TGM1-GFP

injected group compared to the KD 25 injected group (P<0.0048. However, there was

no significant difference between the 5TGM1-GFP and CTRL 25 injected group

(P<0.4667) or between the CTRL 25 and KD 25 groups (P<0.0606).


Figure 6.11: Mice injected with N-cad KD 25 cells had less tumour and reduced tumour take compared to mice injected with CTRL 25 cells, analysed by flow cytometry. 5TGM1-GFP (N=4), CTRL 25 (N=6) and N-cad KD 25 (N=6) cells (2x106) were injected (IV) into 11 week old C57BL/KaLwRij mice. At 25 days P-I (when the first mouse developed hind limb paralysis) all mice were sacrificed. The left and right femora were dissected and the BM was isolated. Flow cytometry was used to quantify the number of GFP positive cells in the BM using a FACS Calibur and Cell quest software. a Representative flow cytometry plots demonstrating the GFP positive tumour in the 5TGM1-GFP (i), CTRL 25 (ii) and KD 25 (iii) injected mice. b. The percentage of GFP positive tumour burden (in R1 of the FACS plot) in mice injected with 5TGM1-GFP, CTRL 25 and KD 25 cells. c. The percentage of tumour take per mouse, in mice injected with 5TGM1-GFP, CTRL 25 and N-cad KD 25 cells. d. The percentage of tumour take per femora, in mice injected with 5TGM1-GFP, CTRL 25 and N-cad KD 25 cells. Data was analysed using a Mann-Whitney U test to compare tumour burden between mice injected with CTRL 25 and N-cad KD 25 cells. A Fisher’s exact test was used to compare the number of mice and femora with positive tumour burden between the 5TGM1-GFP, CTRL 25 and N-cad KD 25 injected mice. Each mouse was a biological replicate.

ai

ii

iii

b

c d

When the number of femora, which were positive for GFP was calculated, each

femora in the 5TGM1-GFP injected mice group was positive for GFP positive (100%),

6 out of 12 femora were positive for GFP in the CTRL 25 groups (50%) and again no

bones were positive for GFP in the KD 25 group (0%), as demonstrated in Figure

6.11d. Using the Fisher’s exact test there were significantly less femora which were

positive in the CTRL 25 (P<0.0445) and KD 25 (P<0.0001) injected mice compared to

the 5TGM1-GFP injected mice. There were also significantly less femora which were

positive for GFP in the KD 25 injected group compared to the CTRL 25 injected group

(P<0.0137).

6.4.11 IHC to determine the effect of knocking-down N-cad in the 5TGM1-GFP cells

upon myeloma cell growth in vivo

Immunohistochemistry was also used to assess tumour burden. Figure 6.11a

demonstrates representative images of histological sections of tibiae stained for

CD138, taken from mice injected with CTRL 25 and N-cad KD 25 injected mice.

Discrete tumour colonies were not observed in any of the KD 25 mice tibiae, though

some CD138 staining was present in some of the bones as shown in Figure 6.12aiii-

iv. However, the staining present was not discrete membranous staining as observed

in the CTRL 25 sections. In addition, all 5TGM1-GFP injected mice had fully infiltrated

bones (data not shown), which correlated with the flow cytometry shown in Figure

6.12.

When tumour colony area and number was analysed using the Image Scope software

there was interestingly, a significant decrease in tumour area in mice injected with

N-cad KD cells compared to CTRL injected mice (P<0.0043) (Figure 6.12b). However,

the 5TGM1-GFP injected mice had fully infiltrated bones with between 80.78-99.77%

(colony area) (data not shown), compared to the low percentage in the CTRL 25

injected mice. There was also a decrease in the number of colonies present in the BM

of mice injected with CTRL 25 or KD 25 cells (Figure 6.12c) but this failed to reach

significance.

When the number of mice which demonstrated positive tumour burden by CD138

staining (positive tumour burden was classed as colony area greater than 0.2%) were

compared between the groups using a Fisher’s exact test, significantly more mice

(expressed as a percentage) in the CTRL 25 group had tumour compared to the KD

25 group (P< 0.0152), shown in Figure 6.12d.

Figure 6.12: Mice injected with N-cad KD 25 cells had significantly less tumour compared to mice injected with CTRL 25 cells, analysed by IHC. Two million CTRL 25 (N=6) and KD 25 (N=6) cells were injected (IV) into 11 week old C57BL/KaLwRij mice. At 25 days P-I (when the first mouse developed hind limb paralysis) all mice were sacrificed. The left tibia was dissected and processed for histology by fixation for 24 hours in 4% PFA, decalcification for 2 weeks in 0.5% PFA/0.5 M EDTA followed by dehydration and processing into wax. Sections were cut (3 µm) and stained by sandwich IHC for CD138. Sections were analysed using an Aperio Scan Scope slide scanner and Image Scope software to determine tumour colony area and number per mm2 of BM. a Representative images of CD138 stained sections of tibiae from mice injected with CTRL 25 cells (i-ii) and KD 25 cells (iii-iv). b. The percentage of total colony area in mice injected with CTRL 25 and KD 25 cells c. The number of tumour colonies present in mice injected with CTRL 25 and KD 25 cells. d. Tumour take in mice injected with CTRL 25 and N-cad KD 25 cells. Data was analysed using a Mann-Whitney U test to compare colony area and number between the groups and a Fisher’s exact test was used to compare the number of mice with tumour between the CTRL 25 and KD 25 injected mice. Each mouse was a biological replicate.

ai

d c b

ii

iii iv


6.4.12 Micro-computed tomography to determine the effect of knocking-down N-

cad in the 5TGM1-GFP cells upon bone parameters in vivo

To assess any differences in bone parameters between the groups µCT was used.

Figure 6.13 demonstrates the percentage of trabecular bone volume (BV/TV),

trabecular number (tb.n) and trabecular thickness (tb.th) for the tibiae of mice

injected with CTRL 25 and KD 25 cells. Statistically, there were no significant

differences in percentage BV/TV, tb.n and tb.th between the CTRL 25 and KD 25

injected mice. However, there did seem to be a slight trend whereby the N-cad KD

25 cell injected mice had slightly higher values for the percentage BV/TV and tb.n,

compared to the CTRL 25 injected mice.

Figure 6.13: There were no significant differences in bone parameters between mice injected with CTRL 25 cells compared to mice injected with N-cad KD 25 cells The right tibiae from mice injected with 2x106 CTRL 25 (N=6) and KD 25 (N=6) cells were dissected, fixed in 10% formalin and analysed by µCT analysis using a SkyScan 1172 at 50 Kv, 200 µA, 0.5 mm aluminium filter and pixel size of 4.3 µm2. Images were reconstructed using N-Recon Software. Trabecular bone 0.2 mm below the growth plate was analysed using Ct-an and Batman software. a. Representative images of cross sectional tibiae taken from mice injected with CTRL 25 cells (i) and KD 25 cells (ii). b. Percentage BV/TV. c. Tb.th. d. Tb.n. Data was analysed using a Mann-Whitney U test (P<0.05) to compare bone parameters between the CTRL 25 and N-cad KD injected mice. Each mouse was a biological replicate.

Lesion number and area were calculated using ImageJ software, as shown in Figure

6.14. There was a trend towards a reduction in lesion number and area in mice

b ai

ii

c

d


injected with the KD 25 cells compared to the CTRL 25 cells however, this did not

reach significance.

Figure 6.14: There were no significant differences in lesion number or area between mice injected with CTRL 25 or N-cad KD 25 cells. The right tibia from each mouse injected with 2x106 CTRL 25 (N=6) and KD 25 (N=6) cells was dissected, fixed in 10% formalin and analysed by µCT analysis using a SkyScan 1172 at 50 Kv, 200 µA, 0.5 mm aluminium filter and pixel size of 4.3 µm2. Images were reconstructed using N-Recon Software and ImageJ software was used to assess lesion number and area. a. Lesion number. b. Lesion area. Data was analysed using a Mann-Whitney U test to determine differences in lesion number and area between mice injected with CTRL 25 and N-cad KD 25 cells. Each mouse was a biological replicate.

6.5 Discussion

The aims of this chapter were to determine the role of HSC niche molecules in MC

adhesion to osteoblastic cells in vitro as well as tumour growth in vivo. To achieve

these aims, lentiviral particle shRNAs were used to KD HSC molecules in the 5TGM1-

GFP cells. Before transduction, optimisation of the technique was conducted to

determine the optimum dose of puromycin and polybrene and also a suitable MOI

of virus using GFP transduction.

Firstly, puromycin concentrations were determined by conducting dose titration

killing curves using TO-PRO-3 and flow cytometry. The final concentration of

b a


puromycin used for the 5TGM1-GFP cells was 5.0 µg/ml and 7.5 µg/ml was used for

the 5T33-WT cells. These concentrations produced significantly more cell death

compared to the control and they were also the lowest doses required to kill the

maximum amount of cells within four days, as suggested by the manufacturer. The

potency of puromycin seemed to vary depending on the cell type, as shown by the

higher dose required to kill the maximum amount of 5T33-WT cells. However, this

higher dose was also used in previous studies where Glick et al (172) used puromycin

at 7.0 µg/ml for the 5T33-WT cells. The difference between the sensitivity of MCs to

puromycin has not previously been reported, but it was interesting that the 5TGM1-

GFP cells seemed to be more sensitive to puromycin compared to the 5T33-WT cells.

Polybrene optimisation was then conducted by assessing differences in cell number

following polybrene treatment after 24 hours. As there was no significant effect upon

cell number following treatment, with the exception of 10.0 µg/ml using the 5TGM1-

GFP cells, 5.0 µg/ml was used as advised by the manufacturer. Further experiments

could have been investigated to determine long-term effects of polybrene treatment

although, the polybrene was only required to be incubated with the cells for 24 hours

and the manufacturer reported no effects of toxicity at the 5.0 µg/ml dose.

Following puromycin and polybrene optimisation, GFP transduction was conducted

in the 5T33-WT cells as a proof of principal that the transduction technique would

work in the MCs. Green fluorescent protein transduction was successful at MOIs of

15 and 25 but not at a MOI of 5 due to a lack of resistance to the puromycin

treatment. This implies that the lentiviral particles were either not successfully

transduced or that the number of puromycin resistance gene sequences integrated

into the host genome using a MOI of 5 was not sufficient to incur resistance to

puromycin, resulting in cell death. Therefore, this suggests that at the higher doses

of virus, an adequate amount of the puromycin resistance gene was integrated into


the genome to inhibit the effect of puromycin on protein synthesis, preventing cell

death.

These optimal conditions were used to KD N-cad and CXCR4 in the 5TGM1-GFP cells.

N-cad was chosen due to its novelty in MM, as previous data is limited. CXCR4 was

chosen as previous studies had largely focussed upon the role of CXCR4 in homing,

but not in the context of a myeloma niche. Previous studies have also used

compounds rather than KDs to block all CXCR4 positive cells, whereas, the KDs used

here, were specific for the MCs.

N-cad was successfully knocked-down at MOIs of 15 and 25 whereas, CXCR4 was

slightly knocked-down at an MOI off 15 (24%) and there was no KD at all using a

higher MOI of 25. From these experiments, it can be concluded that increasing the

MOI did not increase CXCR4 KD. Therefore, increasing the numbers of sequence

insertions into the host’s genome made no difference, implying that the shRNA

sequences used were not effective at knocking down the gene. The lentiviral particles

contained three different shRNA constructs, therefore, three different siRNAs would

bind to three different sequences on the mRNA for CXCR4, increasing the probability

that one sequence would result in the degradation of CXCR4 mRNA. However, in

these experiments these sequences were not effective at knocking-down CXCR4. To

improve this in the future, custom made shRNAs may be required with particular care

to account for splice variants, which do occur within murine CXCR4. It may also help

to use a different promoter such as the U6 rather than the H1, which was reported

to be more effective (188).

As the real-time PCR for N-cad was successful, further validation was conducted using

WB, which confirmed the results found in the real-time PCR. Western blotting

showed that there was a protein KD of 68% and 65% in the 5TGM1-GFP cells at MOIs

of 15 and 25, respectively. However, when the cells were cultured in the absence of


puromycin this resulted in some loss of N-cad KD in the KD 15 cells using both real-

time PCR and WB whereas, using the KD 25 cells the gene and protein KD was stable.

This implies that perhaps the amount of shRNA incorporated into the host DNA at

the MOI of 15 was not sufficient to sustain the KD or that the original selection

process was not selective enough, resulting in the expansion of cells without the

shRNA, therefore, reducing the percentage of population KD.

To confirm any effect upon proliferation in the N-cad KD cells, growth curves were

conducted using the CTRL and N-cad KD cells. There were no significant differences

in proliferation between the CTRL 15 and KD 15 and CTRL 25 and KD 25 cells with

regards to doubling time. This is supported by Groen et al (206) who found that

inhibition of N-cad using an antibody or by knocking-down N-cad in NCI-H929 MCs

had no effect upon proliferation. Whereas, in prostate cancer, antibody inhibition of

N-cad using 1H7 and 2A9 clones resulted in a significant reduction in PC3 and LNCAP-

C1 cell proliferation (203). However, these cells are highly adherent compared to the

MCs, which may influence their requirement for N-cad signalling as a stimulant for

proliferation.

To test the effect of N-cad KD upon MC adhesion to osteoblastic cells, co-culture

experiments were conducted using the conditions from Chapter V. Knock-down of N-

cad in the 5TGM1-GFP did not significantly impair their ability to adhere to primary

OLCs cultured in OGM for 3, 14 and 28 days. There was however, a trend decrease in

adhesion of KD 25 cells to primary OLCS at each of the time points compared to the

CTRL 25 cells. Whereas, there was no obvious difference between the CTRL 15 and

KD 15 cells. The difference between the MOI 15 and 25 cells may be due to the

stability of the KD in the cells. As identified previously, N-cad KD using the MOI of 15

was unstable compared to the MOI of 25 and therefore, the increase in N-cad protein

could remove any inhibitory effect in the adhesion to the primary OLCs. As only a

trend was observed towards a decrease in the adhesion of the KD 25 cells to the


primary OLCs this might imply that the N-cad is not the sole molecule responsible for

MC adhesion to osteoblastic cells. Other molecules previously implicated in this

interaction are; Notch-1 and Jag-1 (described for their role in osteoblast adhesion),

intercellular adhesion molecule (ICAM) 1 (expressed by osteoblastic cells) (272) and

LFA-4 (expressed by MCs) (273) and vascular cell adhesion molecule-1 (VCAM-1)

(expressed by BM stromal cells) (274) and VLA-4 (expressed by MCs) (275).

A lack of significance between N-cad KD and controls may also be due to the

percentage of N-cad KD in the cells. N-cad has not been knocked-out but knocked-

down and using RT-PCR and WB it is uncertain whether 71% of cells are knocked-

down or whether protein within each individual cells is knocked-down by 71%. This

therefore, means that N-cad is still present to a certain extent and therefore, even

low levels may result in adhesion of MCs to the primary OLCs. However, to disprove

this point, a previous paper has shown a significant reduction in the adhesion of N-

cad KD NCI-H929 MCs to C3H10T1/2 cells (206). Although, as the C3H10T1/2 cells are

MSCs as opposed to preosteoblasts or mature osteoblasts, this could be a potential

explanation as to why these results differ from my own. As stated previously in

Chapter V this technique may also not provide reliable results due to elements within

the technique, which may need further optimisation, such as the wash steps, cell

numbers and types of media or FBS used.

An initial pilot study was conducted, firstly, to test whether the CTRL 25 and KD 25

cells, which had been cultured extensively in vitro, still homed and grew in the BM

and secondly, to determine whether there was a difference in tumour burden

between the mice injected with the KD 25 cells compared to those injected with the

CTRL 25 cells. CTRL and KD 25 cells were used due to their high level of gene and

protein KD and low gene and protein KD variability when in puromycin free

conditions compared to the CTRL 15 and KD 15 cells. The flow cytometry data

confirmed that the CTRL 25 and KD 25 cells cultured extensively in vitro did not either


home to the bone as effectively or grow as quickly in vivo compared to the normal

5TGM1-GFP cells. This was shown by higher tumour take and tumour burden in the

5TGM1-GFP injected mice compared to the CTRL 25 and N-cad KD 25 injected

animals. As stated previously this was expected, as the 5TGM1-GFP cells require in

vivo passaging to remain tumourgenic.

The aspect of the data that was of most interest was the difference in tumour take

and tumour burden between the CTRL 25 and KD 25 injected mice by flow cytometry

and IHC. However, as stated previously, this was an initial pilot study and therefore,

may need to be repeated to assess reproducibility. Based on the flow cytometry data

there was a significant reduction in tumour burden in the mice injected with the KD

25 cells compared to those injected with the CTRL 25 cells. When the number of mice

with tumour were compared, 67% of the mice in the CTRL 25 group were positive

compared to 0% of mice injected with KD 25 cells. Although, this difference was not

significant as two mice in the CTRL group did not had have tumour. However, when

the number of bones with tumour rather than per mouse were analysed, (as two

femora per mouse were analysed), significantly fewer femora in the KD 25 injected

group were positive for tumour compared to the CTRL 25 group. In addition, the IHC

staining supported the flow cytometry results in the pilot study. Firstly, full

infiltration of the BM was observed in all of the 5TGM1-GFP injected mice by CD138

staining. In the CTRL 25 injected mice, distinct colonies were visualised whereas, in

the KD 25 injected mice, only small less discrete colonies were observed. This slightly

contradicts the flow cytometry data where no GFP positive tumour cells were present

in the KD 25 injected mice. However, flow cytometry can be limiting, firstly due to

the way the BM is flushed and secondly due to the sensitivity of detecting low

numbers of GFP by the machinery and the software. Despite differences in

techniques, both the flow cytometry and IHC results suggested that there was a

significant reduction in tumour burden in the KD 25 injected mice compared to the

CTRL 25 injected mice and there was also a trend towards fewer tumour colonies

present in the KD 25 injected mice by IHC. These data may potentially suggest a role


for N-cad with regards to MC homing, MC survival in the BM following injection as

well as MC proliferation, though the latter may be questionable as no proliferative

effects were observed in the N-cad KD cells in vitro, though factors in vivo may affect

this.

Micro-CT was also conducted to examine the effect of N-cad KD upon bone disease,

as the 5TGM1-GFP model results in lytic lesions in the mice. When comparing the

CTRL 25 and KD 25 trabecular bone parameters, there were no significant differences

between these groups. However, there was a slight trend towards an increase in the

BV/TV, trabecular number, lesion number and lesion area in the KD 25 injected

animals compared with the CTRL 25 group, which correlated with tumour burden.

However, as little tumour burden was present in the mice it was unsurprising that no

significant differences were found, as there may have not been sufficient tumour

present to induce increases in osteoclastic bone resorption and decreases in

osteoblastic bone formation. In addition, as a naïve mouse group was not used in this

study there was also no comparator for normal bone parameters, therefore, whether

either group of mice had evident bone disease was undetermined.

The data in this chapter, particularly with regards to the pilot in vivo study, provides

evidence in support of the hypothesis that N-cad is important for myeloma

development. From these studies, the precise role of N-cad in myeloma development

is unknown but these experiments may imply that N-cad is involved in MC homing,

survival and/or proliferation; however, due to poor tumour burden and low N

numbers these findings are not completely conclusive. Currently, very little evidence

is available to determine the role of N-cad in MM and no studies have looked at the

effect of knocking N-cad down in MM upon tumour burden at the end-stage of

disease. Groen et al (206) has however, examined the effect of knocking-down N-cad

in NCI-H929 cells upon the number of MCs in the BM after 3 days post-injection and

found that fewer N-cad KD NCI-H929 cells were present in the bone and more of


these cells were in the circulation compared to control NCI-H929 cells. Therefore, N-

cad may be important for homing of MCs to the BM as well as attachment to a MC

niche and MC survival. In addition, Sadler et al (207) found that inhibition of MC

adhesion to stromal cells in vitro using an anti-N-cad antibody resulted in increased

proliferation implying that MC attachment to stromal cells via N-cad is important for

control of cell proliferation and cell cycle. However, this is contradictory to the

experiments conducted in this chapter, which demonstrated reduced TB in mice

injected with N-cad KD cells. However, as stated previously it is unknown as to

whether the reduction in TB is due to reduced MC homing, survival or proliferation

in the BM post-injection, and also as this is a KD not KO, cells arriving in the BM maybe

N-cad positive and so no effect on proliferation would be observed. In addition, the

experiment conducted by Sadler et al was conducted in vitro, and therefore, the

behaviour of these cells in vivo may not be comparable.

Previous studies also determined that N-cad was an important molecule with regards

to cancer cell growth and metastasis as described in the introduction. Tanaka et al

(203) demonstrated that inhibition of N-cad using two different monoclonal

antibodies significantly inhibited PC3 and LAPC9-CR (prostate cell lines) invasion,

proliferation and attachment to fibronectin in vitro and significantly reduced tumour

volume of both PC3 and LAPC9-CR subcutaneous tumours as well as preventing

metastasis to the lymph nodes in vivo. N-cad expression in breast cancer cell lines

including MDA-MB-435, SUM-159 cells and MCF-7 cells was associated with higher

motility and invasion in vitro (204, 268) and increased metastasis in vivo (268). N-cad

expression has also been implicated in melanoma whereby inhibition of N-cad in

melanoma cell lines using an inhibitory antibody reduced adhesion and migration to

fibroblasts in vitro (270).

In conclusion, I have been able to optimise a technique for the successful KD of one

of the HSC niche molecules in the 5TGM1-GFP cells. However, the effect of the KD of


N-cad in the MCs with regards to adhesion to osteoblasts and its effect upon in vivo

are still not completely conclusive. Therefore, further experiments are required to

provide more conclusive evidence for the role for N-cad in an osteoblastic niche in

bone using either the KD cells developed here or an N-cad inhibitor.

Chapter VII: General discussion Page 234

Chapter VII:

General discussion


7.1 General discussion

To date, MM is an incurable disease whereby only 5% of patients survive 5 years post-

diagnosis. It results in severe disorders such as renal failure, lytic bone disease and

susceptibility to infection, which in turn cause morbidity. It has been speculated that

high rates of patient relapse in MM patients are due to a chemoresistant population

of myeloma cells, which are able to survive even after treatment, rendering current

chemotherapeutic strategies ineffective at completely eradicating disease.

Therefore, new treatment strategies are required in an attempt to provide curative

treatments. In addition, preventative strategies are required to stop the progression

of MGUS to symptomatic myeloma and asymptomatic to symptomatic myeloma,

which have a 1% chance within one year post-diagnosis, and 10% chance within five

years post-diagnosis, of progression respectively (276).

Due to the potential survival of chemoresistant cells in patients after treatment, this

led to the hypothesis that MM cells exist in a niche in the BM, which determines their

proliferation, and promotes survival upon treatment with chemotherapeutic agents.

This rationale was based upon previous studies, which identified a HSC niche in the

BM. This HSC niche firstly; demonstrated the requirement of osteoblastic cells for

HSC survival in the BM by a series of studies which reduced (71, 78, 79) or increased

(81, 82) osteoblastic cell number using murine transgenic studies and hormone

treatments. Secondly; HSCs seemed to localise and adhere specifically to osteoblastic

cells (83) as well as endothelial cells (88, 89, 277) in vivo, which seemed a critical

interaction for their survival. In addition, HSCs adhering to osteoblastic cells in the

BM resided in a quiescent state but once activated began to proliferate (73, 142).

To firstly provide evidence of MM cell localisation to osteoblastic cells in a similar

way to HSCs, several in vivo studies were conducted by members of the myeloma

research group at The University of Sheffield and The Garvan Institute, Sydney

Australia (Paper in submission “A Novel Method for Longitudinal Tracking of Dormant


Cancer Cell Activation in the Skeleton as a Platform for Drug Discovery”).

C57BL/KaLwRij mice were injected with 1,1′-dioctadecyl-3,3,3′,3′-

tetramethylindodicarbocyanine perchlorate (DiD) labelled 5TGM1-GFP cells and

after 18 and 72 hours single labelled MCs were found in close proximity to bone.

However, over a time course, the majority of cells moved away from the bone surface

and began to proliferate whereas, only single un-dividing DiD labelled cells, which

were potentially quiescent, were still adjacent to the bone surface. Therefore, these

data may imply that the MCs that were in contact with bone were dormant, whereas

those that were further away from the bone were actively dividing. However,

without staining these cells with cell cycle dyes it is difficult to conclude whether

these cells are actively dividing or quiescent. In support of this work, Chen et al (70)

found that PKH26 labelled MCs were found in close proximity to osteoblastic cells in

the bone and a higher frequency of PKH26 positive cells were present in osteoblastic

niches compared to vascular and splenic niches in vivo. In addition, PKH26 positive

cells were more resistant to chemotherapeutics in vitro compared to PKH26 negative

MCs. Therefore, this supports the hypothesis that MCs reside in osteoblastic niches,

which may result in MC chemoresistance, potentially mediated by cellular

quiescence.

Within my experiments in this thesis, I conducted in vitro experiments to determine

whether the 5TGM1-GFP cells adhered to primary OLCs, as this again would provide

evidence for the adhesion of MCs to osteoblastic cells (Chapter V). A large proportion

of 5TGM1-GFP cells adhered to primary OLCs in vitro and the adhesion was highest

when the OLCs had been differentiated for 3 days and 14 days compared to 28 days.

This suggested that early primary OLCs may be more important for MM cell adhesion

as described by Zhang et al (83) and Xie et al (84) in the HSC niche.

To determine what molecule interactions may result in the adhesion of MM cells to

osteoblastic cells, the expression of the HSC niche molecules and ligands was

analysed using 5T33-WT, 5T33-GFP and 5TGM1-GFP cells. In addition, the expression

of complementary ligands was explored using murine primary OLCs and MC3T3-E1


osteoblastic cells. It was hypothesised that the MM cells expressed the same

repertoire of molecules as the HSCs and that their complementary ligands were

expressed by osteoblastic cells.

Within my studies, I determined the expression of the HSC niche molecules; CXCR4,

Notch-1, Tie-2 and N-cad, and the complementary ligands; CXCL12 and Jag-1 by the

MCs cultured in vitro (Chapter III), collectively using RT-PCR (endpoint and real-time),

flow cytometry and IF. In addition, BM-derived ex vivo 5T33-GFP and 5TGM1-GFP

cells also expressed the HSC niche molecules CXCR4, Notch-1, Tie-2 and

complementary ligands CXCL12 and Jag-1 by flow cytometry, and BM sections

infiltrated with 5TGM1-GFP cells also expressed CXCR4 and N-cad in vivo (Chapter

IV). In addition, the expression of each of these molecules (excluding CXCR4) was

reduced in the BM-derived 5TGM1-GFP at the end-stage of disease compared to the

in vitro cells whereas, the protein frequency of all HSC niche molecules increased in

the BM-derived 5T33-GFP cells compared to the 5T33-GFP cells cultured in vitro.

Therefore, these data suggest that the expression of these molecules may be

important in disease progression.

The expression of these molecules by MCs is supported by a variety of studies. CXCR4

was expressed by MC lines (170, 193-195) and primary patient MCs (195, 221) and

its ligand CXCL12 was also expressed in human MC lines and patient samples (215).

Notch-1 was expressed by several human MC lines (197) and patient MCs (198) and

expression of Jag-1 was demonstrated in patient MC samples (198, 214). Tie-2 was

also present in human MC lines and patient samples (201) as was Ang-1 (201). N-cad

expression was also determined in patient MM cells (205) and the human MC lines

(206, 207).

I also analysed the expression of the complementary ligands CXCL12, Jag-1, Ang-1

and N-cad, as well as the HSC niche molecules CXCR4, Notch-1 and Tie-2 using the

MC3T3-E1 cells and primary OLCs (Chapter III). Each cell type expressed each

complimentary ligand as well as the HSC niche molecule Notch-1 and the primary

OLCs also expressed Tie-2. When the primary OLCs were cultured with OGM (Chapter


V), there were no significant differences in the expression of the HSC niche molecules

or ligands, though there was a trend towards a decrease in Jag-1 and N-cad

expression over-time. Therefore, differentiation of these cells did not significantly

affect HSC niche and ligand expression, though as stated previously without the

suitable controls this is difficult to conclude.

The expression of each of the complementary ligands by osteoblastic cells was

supported by previous studies. CXCL12 was expressed by human and murine

osteoblast cell lines (111, 196); Jag-1 was expressed by human primary BM stromal

cells (199) and the MC3T3-E1 osteoblastic cells (200); Ang-1 was expressed by murine

primary osteoblasts (85); and N-cad was expressed by several stromal and osteoblast

cell lines (206, 208, 219).

Taken together, these studies supported the hypothesis that MCs express the same

repertoire of niche molecules as HSCs and that their complementary ligands are

expressed by osteoblastic cells. However, the function of each of these molecules in

a MM cell niche was unknown. Therefore, to determine the role of these molecules

I attempted to KD both CXCR4 and N-cad in the 5TGM1-GFP cells, with a successful

KD of both gene and protein using N-cad shRNA (Chapter VI). As stated previously,

these molecules were chosen, as little is known about the role of N-cad in MM,

particularly as a candidate molecule in a MC niche. Whereas, CXCR4 has been highly

studied with regards to mobilisation of MCs as well as other cell types, however, it

has not been implicated specifically in an osteoblastic niche in MM. In addition,

previous pre-clinical and clinical studies have focussed on inhibiting CXCR4/CXCL12

signalling using antibodies and inhibitors such as AMD3100, which target all CXCR4

positive cells, however, previous studies have not targeted CXCR4 specifically on the

MCs, which the gene KD would achieve.

I then determined the effect of knocking down N-cad in the 5TGM1-GFP upon

adhesion to primary osteoblastic cells in vitro (Chapter VI). In these experiments a

consistent trend towards a reduction in the adhesion of KD 25 cells to OLCs at all

time-points compared to the CTRL 25 cells was observed, however this did not reach


significance. This trend is supported by other studies which showed that KD of N-cad

significantly reduced MC adhesion to osteoblastic cells however, this was using the

human MC line NCI-H929 and C3H10T1/2 stromal cells (206). My results may imply

that N-cad is not the only molecule responsible for MC adhesion to osteoblastic cells

and as this is a KD of protein rather than KO, residual N-cad expression may still result

in myeloma adhesion to osteoblastic cells.

An initial in vivo pilot study was then conducted to test firstly, whether extensive

growth of the CTRL 25 and KD 25 in vitro would affect their ability to home and grow

in the BM and secondly to determine any differences in tumour burden in mice

injected with CTRL and N-cad KD cells. I hypothesised that the cells that had been

cultured extensively in vitro would not home and/or grow at the same rate as the

standard 5TGM1-GFP cells. I also hypothesised that tumour burden would be

reduced in the mice injected with the KD 25 cells compared to CTRL cells. In support

of these hypotheses, mice that had been injected with either the CTRL 25 or KD 25

cells, had significantly less tumour take (determined by flow cytometry and IHC)

compared to the mice injected with the standard 5TGM1-GFP cells. Interestingly, to

support the second part of the hypothesis, mice injected with the KD 25 cells had no

GFP positive tumour determined by flow cytometry and a significant reduction in

tumour area compared to mice injected with CTRL 25 cells. In addition, mice injected

with the KD 25 cells also had a significantly lower take rate and fewer bones with

tumour present compared to CTRL 25 injected mice using flow cytometry and IHC.

This pilot study therefore, highlights the potential importance of N-cad for myeloma

tumour development in the BM. However, as stated previously these results are not

entirely conclusive due to poor tumour take and low N numbers. These results are

however, supported by previous studies, which have explored the role of N-cad in

other cancers such as breast, prostate and melanoma. In previous studies, N-cad

expression was important for prostate and breast cancer cell invasion and metastasis

(203, 204, 268), and interventions of N-cad inhibited tumour cell growth and

metastasis in prostate and melanoma (203, 278). In addition, Groen et al (206),

demonstrated that N-cad KD in the NCI-H929 MCs resulted in fewer cells homing to

and/or surviving in the BM in vivo, and also found a higher number of these cells


present in the circulation. Therefore, N-cad may potentially be important for trans-

endomigration or adhesion to osteoblastic cells in the BM and subsequent survival.

As this was a pilot study, this experiment needs to be repeated to confirm the results.

In addition, future studies need to focus on the localisation of MCs with the

hypothesis that KD of N-cad will prevent MC attachment to an osteoblastic niche in

the BM, which is required for MC development. In this pilot study, only the latter

time points of active cell growth were determined rather than the initial homing and

colonisation of cells to the bone. Therefore, future studies also need to concentrate

on the early time points of disease to determine differences in the number of cells

originally homing to the bone and also assessing their localisation to osteoblastic cells

at that time.

Other unanswered questions, which this study does not address is the potential for

N-cad positive cells to promote chemoresistance in MCs. In addition, N-cad also

seems to play an important role in epithelial-mesenchymal transition, where the

switching between E-cadherin (E-cad) and N-cad determines the proliferative and

metastatic state of the cells. In prostate cancer, high N-cad expression was associated

with low E-cad expression in prostate cancer biopsies, which correlated with

increasing severity of the prostate cancers (228, 279, 280) which were also associated

with poorer patient outcome. In breast cancer, a reduction of E-cad staining in

patient breast biopsies also correlated with increased severity of disease (281) and,

several breast cancer cell lines which were positive for N-cad but had no evidence of

E-cad expression were more invasive in vitro (204), suggesting again that cadherin

switching may be important for tumour invasion and metastatic potential. Therefore,

this suggests that the relationship between E-cad and N-cad may also be an

important mechanism potentially in MM, where E-cad expression may precede N-

cad expression in quiescent tumour cells after which they may be activated resulting

in cadherin switching from E-cad to N-cad triggering proliferation. Therefore, it would

be of interest to determine whether a cadherin switch occurs throughout the


different stages of myeloma progression and whether it is the initial E-cad positive

cells, which promote quiescence, followed by N-cad expression, which results in

proliferation. In addition, future experiments could also focus on MC survival in vivo

following chemotherapeutic treatment and identifying whether these cells are N-cad

positive and also whether these cells reside in a quiescent state adjacent to

osteoblastic cells, which facilitates resistance. This would then provide the rationale

to inhibit the interaction between the MCs and osteoblastic cells using a potential N-

cad inhibitor, followed by chemotherapy. These are important questions to explore

in the future to further characterise the role of N-cad in MM and a more detailed

plan and further study modifications are discussed in Section 7.2, Future Work.

The majority of functional studies conducted within this thesis focused upon the role

of N-cad in MM rather than the other molecules and ligands. However, it is highly

likely that N-cad is not the sole molecule, which will influence MC adhesion to

osteoblastic cells as well as proliferation in vivo. Clinical trials have previously

explored, or are currently being conducted, to determine the role of CXCR4/CXCL12,

Notch-1/ Jag-1 and Tie-2/ Ang-1 in cancer. Inhibition of CXCR4/CXCL12 signalling has

also been translated into several different clinical trials. For example, inhibition of

CXCR4 using AMD3100 has been used to mobilise HSCs from the BM into the blood

for HSC harvesting in healthy volunteers (282) and has now also been assessed in

phrase three trials in combination with granulocyte colony-stimulating factor (GCSF)

to mobilise HSCs from myeloma patients before receiving chemotherapy (283).

Furthermore, an anti-CXCR4 antibody, known as BMS-936564, inhibited CXCL12

induced migration of B lymphoblast Burkitt lymphoma and T lymphoblast leukaemia

cells in vitro and resulted in a decrease in the volume of subcutaneous tumours using

a myeloma xenograft model (284). The inhibitory peptide, 4F-benzoyl-TN14003, also

known as BKT140, originally used by Menu et al (170) in pre-clinical studies to treat

MM, also inhibited the growth of human non-small cell lung carcinoma cells in vitro

and enhanced chemosensitisation (285) in vivo and reduced the size of subcutaneous

tumours in myeloma xenograft models (286).


Several clinical trials are currently underway, using γ secretase inhibitors such as PF-

03084014 and R04929097, to inhibit Notch-1 signalling. PF-03084014, is currently in

phase I trials for use in patients with grade III breast cancer and T-cell leukaemic

patients. Previous studies have shown a reduction in Notch-1 intracellular signalling

which resulted in inhibition of cell growth and an increase in apoptosis in vitro as well

as a reduction in tumour volume in vivo, using T-cell leukaemic cells (287, 288), as

well as in breast cancer cells (289). In addition, the γ secretase inhibitor RO4929097,

is also currently in phase I clinical trials for use in lymphomas as well as a variety of

metastatic solid tumours. In previous studies, this γ secretase inhibitor also inhibited

tumour growth in human non-small cell lung carcinoma cell xenograft models and in

some cases of solid tumours in humans (290, 291). The effects of these γ secretase

inhibitors are yet to be explored in myeloma but may be of interest in the future.

Arcari et al (292) developed a series of Tie-2 inhibitors some of which reduced

tumour volume in a rat glioblastoma model. Trebananib, a peptide Tie-2 inhibitor has

also been tested in phase I clinical trials. However, use of this peptide did not

considerably reduce progression free survival or overall survival in colorectal cancer

patients (293). Whereas, the Tie-2 inhibitor, CEP-11981, also in clinical trials, reduced

tumour volume and increased survival of animals in both glioblastoma and

melanoma models in vivo (294). The mechanism of tumour reduction in these studies

was not determined but a decrease in angiogenesis was speculated. Therefore, Tie-

2 inhibition could be a future treatment strategy in myeloma, as inhibition may result

in a reduction in angiogenesis, which is increased in myeloma patients (271) and this

may result in a reduction in tumour load.

From the data collected in these studies there are several limitations, which need to

be considered. Firstly, all of the experiments were conducted using murine MCs as

opposed to human cells. The 5TGM1-GFP model has several advantages including

excellent tumour take rate and growth in the BM of C57BL/KalwRij mice as well as

producing osteolytic disease. This model is also fairly representative of patient


myeloma, particularly concerning tumour infiltration and bone disease, though it is

more aggressive than MM disease occurring in patients. Another approach to

investigate MM in pre-clinical studies is to use xenograft models of MM where,

human cell lines and primary MM samples are injected into SCID mice (157-162, 165,

295). I could therefore, in future experiments use human cell lines and patients

samples to assess the expression of the HSC niche molecules and ligands and to

explore the importance of N-cad in vivo in xenograft models of MM. However, these

models also have limitations in that immune-compromised mice are used as opposed

to mice with an intact immune system and again isn’t completely representative of

human disease.

In addition, as mentioned previously, the use of KD cells as opposed to clinically

relevant inhibitors are another limitation. Knock-down of genes to determine the

role of the corresponding protein is a good proof of principal experiment. However,

from these experiments it seemed to produce quite high variability in both the in

vitro KD experiments and in the in vivo experiment. This was possibly due to the

percentage of KD, which was not 100% as well as extensive cell growth in vitro and

antibiotic selection. Knock-down experiments also inevitably will require

interventions using compounds or antibodies. Therefore, for future experiments it

may also be desirable to use peptides as well as functional inhibitory antibodies.

Currently, an N-cad inhibitory peptide, known as adherin-1 (ADH-1), is in clinical trials

for use in treating patients with solid tumours. Adherin-1 is a cyclic peptide with a

specific “HAV” amino acid sequence which inhibits N-cad and its function was

originally determined within retinal explants, resulting in a reduction in neurite

outgrowth and a reduction in adhesion of LN-229 cells (a glioblastoma cell line) to

recombinant N-cad in vitro (296). The use of this peptide was particularly beneficial

in rats treated with melanoma DM366 cells by subcutaneous injection, followed by

ADH-1 and melphalan treatment. Rats with melanoma, when treated with ADH-1 and

melphalan in combination, had significantly lower tumour volume compared to those

treated with ADH-1 and melphalan alone (278). Again this compound may be a

suitable therapeutic in MM.


In conclusion, these experiments provide evidence in support of the overall

hypothesis, that “myeloma cells exist within an osteoblastic niche similar to that

occupied by HSCs”. Within these experiments, MCs expressed the same repertoire of

molecules expressed by HSCs, and their complementary ligands were expressed by

osteoblastic cells at varying stages of differentiation. In addition, the KD of one of

these key molecules, N-cad, reduced (though not significantly) the adhesion of MCs

to osteoblastic cells in vitro and potentially reduced tumour growth in vivo. This

reduction may be due to fewer cells originally homing to the BM or inhibition in the

adhesion of MCs to osteoblastic cells, which may be an essential interaction for MC

survival in the BM. However, further experiments are required to confirm the role of

N-cad in myeloma development and to determine the precise role of osteoblastic

cells in this interaction. These studies do however; provide the rationale to continue

to target the niche with the theory that detaching cells from the niche will result in

chemosensitisation, which can be targeted by conventional chemotherapies.

However, it is currently unknown if this would in fact increase tumour growth (as

seen by Sadler et al (207) in vitro), and so one would have to proceed with caution

by extensively testing such treatment strategies in preclinical models. In addition, it

is also unlikely that one molecule alone will facilitate MC attachment to a niche in

the BM and in future experiments using a combination of molecule interventions will

be required. Therefore, future experiments aim to establish treatment strategies,

which could more effectively kill MCs and prevent chemo-evasion.

7.2 Future studies

To assess the effect of knocking-down N-cad in the 5TGM1-GFP cells on homing,

localisation and growth over-time in vivo

To further assess the effect of knocking-down N-cad in the 5TGM1-GFP cells upon

tumour growth in vivo, certain aspects of the pilot study need to be repeated and

additional aspects need to be investigated. Firstly, it would be desirable to determine

whether different numbers of 5TGM1-GFP cells home to the BM. This could be done


by labelling the 5TGM1-GFP cells with the membrane dye DiD prior to injection.

Three days after injection, labelled 5TGM1-GFP cells in the BM or circulation could

be quantified by flow cytometry and their location could be determined using multi-

photon microscopy. Secondly, analysis at the end stage of disease would be

repeated. However, rather than sacrificing the mice at 3 weeks P-I the experiment

would continue until the first mouse in either the CTRL 25 or KD 25 group became ill,

as the standard 5TGM1-GFP group would no longer be required. Tumour burden

would be quantified using IHC and flow cytometry and as the length of the study

would be extended this would allow larger tumours to form in the CTRL 25 group,

which may reduce variability in the tumour burden analysis. Another endpoint would

be to analyse the number of single DID labelled cells present in the bones of mice at

the end stage of disease and to determine the proliferative state of these cells to

establish whether they are quiescent. To reduce variability, mouse numbers in the

groups would be increased. This would be based on power calculations using

standard deviations from the pilot study. A PBS control group would also be used to

provide background CD138 positive plasma cell analysis to quantify CD138 staining

at the early time points of disease by IHC as well as providing basal trabecular bone

parameter values to determine any differences between the PBS control group and

the CTRL 25 and KD 25 injected mice.

Determining the effects of a clinically relevant inhibitor of N-cad upon myeloma

cell homing, colonisation and subsequent growth in the bone marrow

The N-cad KD experimental design could also be used to determine the effects of a

clinically relevant inhibitor of N-cad upon MC homing, colonisation and subsequent

growth in the BM. Initially, optimisation experiments to determine a suitable dose of

an inhibitor, such as ADH-1 or inhibitory antibodies, would be required to assess

toxicity and whether the compounds have any effect on tumour burden at the end

stage of disease. If differences were seen, it would then be desirable to analyse the

early stages of disease, using DID labelled MCs to establish whether treatment of

mice with an N-cad inhibitor prevented MC homing to the BM and also to determine

the localisation of the MCs in the BM in vivo. A potential limitation to this study would

be that the detachment of MCs from the niche might induce proliferation of these


“dormant” cells increasing tumour burden as well as the potential to increase

metastasis.

Quantifying and profiling the myeloma cells present in the bone marrow following

chemotherapy

To assess whether N-cad or the other HSC niche molecules are important with

regards to MC chemoresistance, further in vitro and in vivo studies are required.

Firstly in vitro chemo-titration studies could be conducted to examine the effect of

different doses upon MC death, using modern chemotherapies such as lenalidomide

or pomalidomide. This would be followed by treatment of cells with optimised doses

of chemotherapies followed by genetic profiling of cells before and after therapy to

determine any key molecules, which may enable chemoresistance.

This would then be translated into in vivo experiments, to establish the genetic

profile of the cells surviving after chemotherapeutic treatment in vivo (identified by

DID labelling), which would include analysing the expression of the HSC niche

molecules. In addition, molecules involved in cell signalling pathways could also be

assessed in the chemoresistant cells to determine whether any of these pathways

are activated following chemotherapeutic intervention.

Determining the effect of N-cad intervention in combination with

chemotherapeutic treatment

Following on from both the separate N-cad intervention studies and chemo-titration

experiments, both of these could be combined to form a new treatment strategy.

Mice could be injected with 5TGM1-GFP cells and upon colonisation at approximately

10 days P-I; they could then be treated with the N-cad inhibitor (inhibitory peptide

or antibody), which may potentially dislodge the attachment of MCs from BM niches,

rendering them more chemosensitive. This could then be followed by

chemotherapeutic intervention using modern agents such as lenalidomide or

pomalidomide as well as potentially in combination with proteasome inhibitors such


as bortezomib or carfilzomib. Following the combination treatment, the mice could

be maintained in a survival study to determine whether mice, which received the

combination treatment, survive longer or do not relapse compared to mice receiving

chemotherapy alone. At the end of these studies the long bones could then be

analysed for DID labelled cell number analysis as well as CD138 IHC to determine

whether any MCs had remained in the BM following chemotherapy or N-cad

intervention treatments. As already discussed, this strategy could be quite unsafe

due to detachment of cells from the niche, which may result in cell proliferation and

an increase in metastasis. However, if the chemotherapeutic strategy is at the correct

dose these cells should be targeted.

Appendices Page 248

Appendices

Appendix 1: Equipment, materials and reagents

1.1 Tissue culture

Table 1.1: Equipment, materials and reagents required for tissue culture

Item Supplier

Disposable tips (10, 200 & 1000 µl) Star labs Eppendorph tubes (0.5 ml & 1 ml) Sarstedt

Flasks (T25, T75 & T175) Nalgene Nunc Ltd Falcon tubes (50 ml) BD Pharmingen

Universal tubes (20 ml) Costar 15 ml tubes BD Pharmingen Bijoux tubes Costar

Petri dishes (60 and 100 mm) Nunc Multi-well plates (6, 12, 24, 48, 96) Nalgene Nunc Ltd.

Cryovials (1.5 ml) Nalgene Nunc Ltd Cell scrapers Corning

Syringes BD Pharmingen MEMα medium (GlutaMAX and phenol red) Gibco

MEMα-Nuc medium (GlutaMAX, deoxyribonucleosides, ribonucleosides and

phenol red) Gibco

RPMI 1640 medium (GlutaMAX and Phenol red)

Gibco

FBS Gibco NEAA (100X) Gibco

NaP (100mM of sodium pyruvate) Gibco PS (10,000 units of penicillin and 10,000

µg/ml of streptomycin) Gibco

FZ (250 µg of Amphotericin B and 205 µg of sodium deoxycholate/ml)

Gibco

PBS Gibco Trypsin/ EDTA (2.5 g trypsin, 0.38 g EDTA and 4Na per litre of Hank’s balanced salt

solution per litre) Gibco

Trypan blue Sigma Aldrich DMSO Sigma Aldrich 70 IMS Fisher Scientific

Haemocytometer Neubauer Cover-slips Fisher Scientific

Appendices Page 249

1.2 Animal cell isolation

Table 1.2: Equipment, materials and reagents required for animal cell isolation

Item Supplier

HBSS (calcium and magnesium) Gibco Collagenase Type II Gibco

EDTA Fluk analytical Insulin Myjector needles (27

Gauge) Terumo

Vacutainers BD Pharmingen RBC lysis (10X) and wash buffer

(10X) R & D Systems

DMEM media containing Gibco Pentabarbitone (concentration) J M Loveridge Ltd

1.3 RNA extraction

Table 1.3: Materials and reagents required for RNA extraction

Item Supplier

Ultra spec (A 14 M guanidine salt/urea RNA isolation system)

BioTecx

Molecular grade chloroform Sigma Aldrich Molecular grade 2-propanol Sigma Aldrich

Molecular grade ethanol Sigma Aldrich DEPC water Ambion RNase Zap Ambion

1.4 Reverse transcription

Table 1.4: Equipment, materials and reagents required for reverse transcription

Item Supplier

0.5 ml RNase and DNase free tubes Alpha laboratories PTC 200 thermal cycler MJ Research

UV hood Jencons Nuclease free water Ambion

Random primers (500 ng/ µl) Promega RNasin (40 U/µl)

Promega

SuperScript III first-strand synthesis system, Reverse transcriptase (200 U µl) Life Technologies

DTT (100 mM) Life Technologies

First strand buffer (5X solution containing 250 mM Tris-HCl (pH 8.3), 375 mM potassium

chloride (KCl), 15 mM MgCl2) Life Technologies

dNTPs (100 mM), Bioline

Appendices Page 250

1.5 Endpoint PCR

Table 1.5: Materials and reagents required for endpoint PCR

Item Supplier

RT-PCR H2O Ambion MgCl2 (50 mM) Bioline

Ammonium reaction buffer (10X)

Bioline

Taq polymerase (5 U/µl) Bioline

Murine PCR primers (5 µM) Thermo Scientific and Life

Technologies

1.6 Gel electrophoresis

Table 1.6 Equipment, materials and reagents required for gel electrophoresis

Item Supplier

Electrophoresis tank Biorad PowerPac Biorad Gel combs Biorad

Gel casting mould Biorad Trizma base Sigma Aldrich

Hyperladder IV 100-1000 BP Bioline Loading buffer (5X) Bioline Ultra-pure agarose Sigma Aldrich Glacial acetic acid Fisher Scientific

Ethidium bromide (10 mg/ml) Life Technologies

1.7 Real-time PCR

Table 1.7: Equipment, materials and reagents required for real-time PCR

Item Supplier

384 well plates Greiner Bio-one Ltd Micro-seal adhesive covers Bio-Rad

7900 Ht real-time PCR machine Life Technologies Endogenous control assays and

experimental probes (20X) Life Technologies

Real-time PCR Master Mix: (2x containing AmpliTaq GoldDNA

Polymerase, Uracil DNA Glycosylase, dNTPs

deoxyuridine triphosphate (dUTP)

Life Technologies

Appendices Page 251

1.8 Flow cytometry

Table 1.8: Equipment, materials and reagents required for flow cytometry

Item Supplier

FACS Calibur machine BD Pharmingen 96 V bottom well plate Sterilin Flow cytometry tubes Elkay Laboratory Products

Cell strainers BD Pharmingen Goat serum Vector Laboratories

BSA Sigma Aldrich Cell dissociation solution Sigma Aldrich Fixation buffer (4% PFA) BD Pharmingen Permeablisation buffer

(Saponin) (10X) BD Pharmingen

Fluorescent antibodies Various suppliers-see Table 2.23 PI (1 mg/ml) Life Technologies

TO-PRO-3 (1 mM) Life Technologies

1.9 Immunofluorescence

Table 1.9: Equipment, materials and reagents required for immunofluorescence

Item Supplier

Hydrophobic barrier pen Vector Laboratories Super-frost slides VWR International

Cover-slips (22 x 22 mm) Menzel glaser Cytospin equipment and

machine Shandon

Hot plate Thermo Scientific Leica AF6000 microscope Leica Microsystems

PBS tablets Oxoid Paraformaldehyde Fisher Scientific

NaOH Fisher Scientific Prolong gold anti-fade aqueous

mounting media with 4’,-6-diamidino-2-phenylindole

(DAPI)

Life Technologies

Tween Fisher Scientific Nail varnish Boots Pharmacy

Appendices Page 252

1.10 IHC

Table 1.10: Equipment, materials and reagents required for IHC

Item Supplier

Processing cassettes Leica Microsystems Processing carousal Leica

TP2010 Leica Microsystems

Blocking System II embedder Raymond Lamb Leica RM 2265 microtome Leica Microsystems

Hot plate Raymond Lamb Water baths Raymond Lamb

Oven Scientific Laboratory Supplies Processing wax

Xylene Fisher Scientific Alcohols Fisher Scientific

EDTA Fisher Scientific Trypsin antigen retrieval kit A.Menarini Diagnostics

H2O2 VWR Streptavidin horse radish

peroxidase Dakocytomation

DPX mountant VWR international Primary antibodies- see table

2.12 Various suppliers- see table

2.12 Goat anti-rat biotinylated

secondary antibody Santa Cruz Biotechnology

Goat anti-rabbit biotinylated secondary antibody

Vector Laboratories

DAB peroxidase substrate kit Vector Laboratories Gill’s haematoxylin Fisher Scientific

Appendices Page 253

1.11 Western blotting

Table 1.11: Equipment, materials and reagents required for Western blotting

Item Supplier

Mini protean WB equipment: gel casting stands, casting frames, electrophoresis tank and electrode assembly, transfer cassettes,

transfer electrodes

Bio-Rad

1.5 mm separator and cover plate Bio-Rad 1.5 mm combs (15 wells) Bio-Rad

Filter paper Whatman Teflon pads Bio-Rad

Tris HCl Sigma Aldrich BCA Sigma Aldrich NaCl Sigma Aldrich

Glycine Fisher Scientific NP-40 substitute Fuka Biochemika

TEMED Sigma Aldrich Laemmli sample buffer (4X) Biorad

PIC Calbiochem Acrylamide National Diagnostics

APS Bioproducts ltd β-mercaptoethanol Sigma Aldrich

SDS Sigma Aldrich Milk powder Cooperative supermarket

Goat anti-rabbit horse radish peroxidase antibody

Dako cytomation

Anti-biotin antibody HRP Cell signalling Rabbit anti β-actin antibody Abcam

Pre-stained all Blue protein ladder Bio-rad Biotinylated molecular weight standards Cell signalling

Nitrocellulose membranes Amersham Methanol Fisher Scientific

Luminol reagent (ECL) Santa Cruz Biotechnology Luminol reagent (ECL) Thermo Scientific

Hyper film Amersham Developing solutions AGFA

Appendices Page 254

1.12 Primary osteoblast lineage cell differentiation

Table 1.12: Equipment, materials and reagents required for primary osteoblast lineage cell differentiation

Item Supplier

Spectra Max M5e plate reader Molecular devices BGP Sigma Aldrich Asc Sigma Aldrich

PNPP tablets, containing PNPP (20 mg/ml) and tris buffer tablets (4 M)

Sigma Aldrich

Alizarin red Sigma Aldrich Quant-it Pico green reagent: containing

the dsDNA picogreen component, tris/EDTA buffer (20X) containing 200 mM Tris-HCl, 20 mM EDTA pH 7.5 and

the lambda DNA standards (100 µg/ml)

Life Technologies

1.13 Gene knock-down using short hair-pin RNA

Table 1.13: Materials and reagents required for short hair-pin RNA knock-down

Item Supplier

Puromycin (25 mg/ml) Santa Cruz Biotechnology Polybrene (10 mg/ml) Santa Cruz Biotechnology

CXCR4 and N-cad shRNA lentiviral particles

Santa Cruz Biotechnology

GFP lentiviral particles Santa Cruz Biotechnology Scrambled control lentiviral

particles Santa Cruz Biotechnology

1.14 Micro-computed tomography

Table 1.14: Equipment, materials and reagents required for micro-computed tomography

Item Supplier

SkyScan 1172 µCT machine SkyScan Sodium dihydrogen

orthophosphate dihydrate VWR international

Disodium hydrogen orthophosphate dihydrate

VWR international

Concentrated formaldehyde (i.e. 37-41%)

VWR international

Appendices Page 255

Appendix 2: Additional information for methodologies

2.1 RNA quantification and contamination analysis

Figure 2.1: An example of a RNA spectrophotometry curve to detect contaminants. Ribonucleic acid was extracted and analysed using a NanoDrop 2000 and NanoDrop 2000 software. A graph was generated which summarised the quality of the RNA and generated contamination ratios (260:280, 260:230).

2.2 RNA fragment length analysis

Figure 2.2: An example of RNA fragment length analysis. Ribonucleic acid integrity was measured using an Agilent Bioanalyser 2100 and analysed using 2100 Expert Software. The RNA integrity was scored and used if the RNA integrity was between 8 and 10.

Appendices Page 256

2.3 Endpoint PCR primer information

Table 2.1: Primer annealing temperatures, cycle number and product size for endpoint PCR

Gene Primer annealing temperature (°C)

Cycle number Product size (BP)

GAPDH 53 25 354 CXCR4 48 35 711 CXCL12 48 35 205 Notch-1 48 35 348

Jag-1 48 35 370 Tie-2 48 35 648 Ang-1 48 35 401 N-cad 55 35 560 CD138 48 35 559

2.4 Endpoint PCR primer sequence information

Table 2.2: Endpoint PCR primer sequences, gene accession numbers and primer melting temperatures

Primer Primer annealing Sequence Accession number

(NM)

Tm (°C)

Sense GAPDH 5’-TTGTCAGCAATGCATCCTGC-3’ 008084 53 Anti-sense GAPDH 5’-GCTTCACCACCTTCTTGATG-3’ 008084 53

Sense CXCR4 5-’ATGGAACCGATCAGTGTGAG-3’ 009911 49

Anti-sense CXCR4 5-‘CTTGGAGTGTGACAGCTTAG-3’ 009911 45

Sense CXCL12 5’-TCAGTGACGGTAAACCAGTC-3’ 021704 47

Anti-sense CXCL12 5’-GCTTTCTCCAGGTACTCTTG-3’ 021704 46

Sense Notch-1 5’-TAATGAGTGCAGCCAGAACC-3’ 008714 50

Anti-sense Notch-1 5’-TCCTCTGTACAGTACTGACC-3’ 008714 41

Sense Jag-1 5’-AGATCCTGTCCATGCAGAAC-3’ 013822 49

Anti-sense Jag-1 5’-ATCATGCCTGAGTGAGAAGC-3’ 013822 49

Sense Tie-2 5’-GTTCGAGGACAGGCTATAAG-3’ 013690 46

Anti-sense Tie-2 5’-CTCTTCATTGGTACAGTGGC-3’ 013690 47

Sense Ang-1 5’-CATTCTTCCAGAACACGACG-3’ 009640 51

Anti-sense Ang-1 5’-GGAGAAGTTGCTTCTCTAGC-3’ 009640 45

Sense N-cad 5’-CAGGTAGCTGTAAACCTGAG-3’ 007664 55 Anti-sense N-cad 5’-CCATTCAGGGCATTTGGATC-3’ 007664 55

Sense CD138 5’-TGTTCCTCCTGAAGATCAGG-3’ 011519 49 Anti-sense CD138 5’-TGATTGGCAGTTCCATCCTC-3’ 011519 52

Appendices Page 257

2.5 Probes used for real-time PCR

Table 2.3: Probes used for real-time PCR

Gene Reference number

β2M Mm00437762_m1

HPRT Mm00446968_m1

TFRC Mm00441941_m1

Col1A1 Mm00801666_g1

Runx-2 Mm00501584_m1

CXCR4 Mm01292123_m1

CXCL12 Mm00445552_m1

Notch-1 Mm00435249_m1

Jag-1 Mm00496902_m1

Tie-2 Mm00443243_m1

Ang-1 Mm00456503_m1

N-cad Mm01162497_m1

CD138 Mm00448918_m1

Appendices Page 258

2.6 Real-time PCR probe efficiencies

Figure 2.3: Primer efficiencies for fluorescent probes used for real-time PCR: Primer efficiencies were determined by adding a serial dilution of normal murine BM cDNA to the fluorescent probes for β2M (a), HPRT (b), TFRC (c) (HK genes), CXCR4 (d), CXCL12 (e), Notch-1 (f), Jag-1 (g), Tie-2 (h), Ang-1 (i), N-cad (j) and CD138 (k), using real-time PCR. The Ct values (y-axis) were plotted against the dilution of cDNA (x-axis) and the slope and R2 values were calculated using linear regression.

a b c

e f

g h i

j k

d

Appendices Page 259

Table 2.6: The percentage of fluorescent probe efficiency

The percentage of fluorescent probe efficiency was calculated using the equation: Efficiency = 10(-1/slope) -1 * 100, using the slope from the graphs calculated in Appendix Figure 2.3. Probes within a range of 90-110% were used in real-time PCR.

2.7 Optimisation experiments to determine the optimum amount of protein required to visualise N-cad using Western blotting

Figure 2.4: Western blotting to optimise the amount of protein used to detect N-cad in 5TGM1-GFP cells. Increasing amounts of 5TGM1-GFP protein (5, 10, 20, 40 and 60 µg) were loaded and run by SDS-PAGE and blotted using an anti-N-cad primary antibody followed by anti-rabbit HRP conjugated secondary Blots were visualised using ECL purchased from Santa Cruz.

Gene Primer efficiency (%)

β2M 93.12 HPRT 107.55 TFRC 72.66

CXCR4 100.16 CXCL12 90.57 Notch-1 94.50

Jag-1 93.09 Tie-2 99.18 Ang-1 98.84 N-cad 98.89 CD138 90.66

Appendices Page 260

2.8 Alkaline phosphatase analysis

Table 2.6: An example of alkaline phosphatase calculations used to determine differences in primary osteoblast lineage cell differentiation

References Page 261

References

1. Sheridan CA. Multiple myeloma. Seminars in Oncology Nursing.

1996;12(1):59-69.

2. Streetly MJ. Myeloma and MGUS. Medicine. 2013;41(5):295-8.

3. Koura DT, Langston AA. Inherited predisposition to myeloma. Therapeutic

Advances in Hematology. 2013.

4. Anderson KC, Alsina M, Bensinger W, Biermann JS, Cohen AD, Devine S, et al.

Multiple myeloma, version 1.2013. Journal of the National Comprehensive Cancer

Network. 2013;11(1):11-7.

5. Bird JM, Owen RG, D’Sa S, Snowden JA, Pratt G, Ashcroft J, et al. Guidelines

for the diagnosis and management of multiple myeloma 2011. British journal of

haematology. 2011;154(1):32-75.

6. Baker A, Braggio E, Jacobus S, Jung S, Larson D, Therneau T, et al. Uncovering

the biology of multiple myeloma among African Americans: a comprehensive

genomics approach. Blood. 2013;121(16):3147-52.

7. DeSantis C, Naishadham D, Jemal A. Cancer Statistics for African Americans,

2013. CA: a cancer journal for clinicians. 2013.

8. Calame K. Plasma cells: finding new light at the end of B cell development.

Nature immunology. 2001;2(12):1103-8.

9. Di Noia J, Neuberger M. Molecular mechanisms of antibody somatic

hypermutation. Biochemistry. 2007;76(1):1.

References Page 262

10. Goldsby RA, Kindt T.J, Osborne B. A. Kuby Immunology. 2nd ed. New York:

Freeman and Company; 2000 2000. 1 p.

11. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple

myeloma. Nature Reviews Cancer. 2012;12(5):335-48.

12. Hallek M, Leif Bergsagel P, Anderson K. Multiple myeloma: increasing

evidence for a multistep transformation process. Blood. 1998;91(1):3.

13. Seidl S, Kaufmann H, Drach J. New insights into the pathophysiology of

multiple myeloma. The Lancet Oncology. 2003;4(9):557-64.

14. Mitsiades C, Mitsiades N, Munshi N, Anderson K. Focus on multiple myeloma.

Cancer cell. 2004;6(5):439.

15. Hoffbrand A.V M, P.A.H, Pettit J.E. Essential Haematology. Fifth ed. Oxford:

Blackwell Publishing LTD; 2006. 1 p.

16. Korbet SM, Schwartz MM. Renal Disease in Multiple Myeloma. 2006.

17. Wirk B. Renal failure in multiple myeloma: a medical emergency. Bone

marrow transplantation. 2011;46(6):771-83.

18. Raje N, Roodman GD. Advances in the biology and treatment of bone disease

in multiple myeloma. Clinical Cancer Research. 2011;17(6):1278-86.

19. Tabernero D, San Miguel JF, Garcia-Sanz M, Najera L, García-Isidoro M, Perez-

Simon JA, et al. Incidence of chromosome numerical changes in multiple myeloma:

fluorescence in situ hybridization analysis using 15 chromosome-specific probes. The

American journal of pathology. 1996;149(1):153.

20. Rajkumar SV. Multiple myeloma: 2012 update on diagnosis, risk‐stratification,

and management. American journal of hematology. 2012;87(1):78-88.

References Page 263

21. Rajan AM, Rajkumar SV. Diagnostic Evaluation of Monoclonal Gammopathy

of Undetermined Significance. European journal of haematology. 2013.

22. Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, et al.

Monoclonal gammopathy of undetermined significance (MGUS) consistently

precedes multiple myeloma: a prospective study. Blood. 2009;113(22):5412-7.

23. Waxman AJ, Kuehl M, Balakumaran A, Weiss B, Landgren O. Smoldering

(asymptomatic) multiple myeloma: revisiting the clinical dilemma and looking into

the future. Clinical Lymphoma Myeloma and Leukemia. 2010;10(4):248-57.

24. Smith A, Wisloff F, Samson D. Guidelines on the diagnosis and management

of multiple myeloma 2005. British journal of haematology. 2006;132(4):410-51.

25. Kyle R, Child J, Anderson K, Barlogie B, Bataille R, Bensinger W, et al. Criteria

for the classification of monoclonal gammopathies, multiple myeloma and related

disorders: a report of the International Myeloma Working Group. British journal of

haematology. 2003;121(5):749-57.

26. Durie B, Salmon S. A clinical staging system for multiple myeloma correlation

of measured myeloma cell mass with presenting clinical features, response to

treatment, and survival. Cancer. 1975;36(3):842-54.

27. Paget J. The distribution of secondary growth in cancer of the throat. Lancet.

1889;1:571.

28. Rhodin J. Histology: a text and atlas: Oxford University Press New York; 1974.

29. Teitelbaum SL. Bone resorption by osteoclasts. Science.

2000;289(5484):1504-8.

References Page 264

30. Pivonka P, Zimak J, Smith D, Gardiner B, Dunstan C, Sims N, et al. Model

structure and control of bone remodeling: A theoretical study. Bone. 2008;43(2):249-

63.

31. Proff P, Römer P. The molecular mechanism behind bone remodelling: a

review. Clinical oral investigations. 2009;13(4):355-62.

32. Croucher P, Shipman C, Lippitt J, Perry M, Asosingh K, Hijzen A, et al.

Osteoprotegerin inhibits the development of osteolytic bone disease in multiple

myeloma. Blood. 2001;98(13):3534.

33. Galson DL, Silbermann R, Roodman GD. Mechanisms of multiple myeloma

bone disease. BoneKEy reports. 2012;1(8).

34. Sezer O, Heider U, Jakob C, Zavrski I, Eucker J, Possinger K, et al.

Immunocytochemistry reveals RANKL expression of myeloma cells. Blood.

2002;99(12):4646-7.

35. Han J, Choi S, Kurihara N, Koide M, Oba Y, Roodman G. Macrophage

inflammatory protein-1 {alpha} is an osteoclastogenic factor in myeloma that is

independent of receptor activator of nuclear factor {kappa} B ligand. Blood.

2001;97(11):3349.

36. Oyajobi BO, Franchin G, Williams PJ, Pulkrabek D, Gupta A, Munoz S, et al.

Dual effects of macrophage inflammatory protein-1α on osteolysis and tumor burden

in the murine 5TGM1 model of myeloma bone disease. Blood. 2003;102(1):311-9.

37. Filella X, Blade J, Guillermo A, Molina R, Rozman C, Ballesta A. Cytokines (IL-

6, TNF-alpha, IL-1alpha) and soluble interleukin-2 receptor as serum tumor markers

in multiple myeloma. Cancer detection and prevention. 1995;20(1):52-6.

References Page 265

38. Yamamoto I, Kawano M, Sone T, Iwato K, Tanaka H, Ishikawa H, et al.

Production of interleukin 1β, a potent bone resorbing cytokine, by cultured human

myeloma cells. Cancer research. 1989;49(15):4242-6.

39. Lee JW, Chung HY, Ehrlich LA, Jelinek DF, Callander NS, Roodman GD, et al. IL-

3 expression by myeloma cells increases both osteoclast formation and growth of

myeloma cells. Blood. 2004;103(6):2308-15.

40. Sati HI, Apperley JF, Greaves M, Lawry J, Gooding R, Russell RGG, et al.

Interleukin‐6 is expressed by plasma cells from patients with multiple myeloma and

monoclonal gammopathy of undetermined significance. British journal of

haematology. 1998;101(2):287-95.

41. Yaccoby S, Wezeman MJ, Henderson A, Cottler-Fox M, Yi Q, Barlogie B, et al.

Cancer and the Microenvironment Myeloma-Osteoclast Interactions as a Model.

Cancer Research. 2004;64(6):2016-23.

42. D'Souza S, Kurihara N, Shiozawa Y, Joseph J, Taichman R, Galson DL, et al.

Annexin II interactions with the annexin II receptor enhance multiple myeloma cell

adhesion and growth in the bone marrow microenvironment. Blood.

2012;119(8):1888-96.

43. Abe M, Hiura K, Wilde J, Shioyasono A, Moriyama K, Hashimoto T, et al.

Osteoclasts enhance myeloma cell growth and survival via cell-cell contact: a vicious

cycle between bone destruction and myeloma expansion. Blood. 2004;104(8):2484-

91.

44. Moreaux J, Cremer FW, Reme T, Raab M, Mahtouk K, Kaukel P, et al. The level

of TACI gene expression in myeloma cells is associated with a signature of

microenvironment dependence versus a plasmablastic signature. Blood.

2005;106(3):1021-30.

References Page 266

45. Boyle W, Simonet W, Lacey D. Osteoclast differentiation and activation.

Nature. 2003;423(6937):337-42.

46. Nakamura M, Udagawa N, Matsuura S, Mogi M, Nakamura H, Horiuchi H, et

al. Osteoprotegerin regulates bone formation through a coupling mechanism with

bone resorption. Endocrinology. 2003;144(12):5441-9.

47. Seidel C, Hjertner Ø, Abildgaard N, Heickendorff L, Hjorth M, Westin J, et al.

Serum osteoprotegerin levels are reduced in patients with multiple myeloma with

lytic bone disease. Blood. 2001;98(7):2269-71.

48. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, Barlogie B, et al. The role of the

Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple

myeloma. The New England journal of medicine. 2003;349(26):2483.

49. Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy JD, Evans HR, et al.

Inhibiting Dickkopf‐1 (Dkk1) Removes Suppression of Bone Formation and Prevents

the Development of Osteolytic Bone Disease in Multiple Myeloma. Journal of Bone

and Mineral Research. 2009;24(3):425-36.

50. Oshima T, Abe M, Asano J, Hara T, Kitazoe K, Sekimoto E, et al. Myeloma cells

suppress bone formation by secreting a soluble Wnt inhibitor, sFRP-2. Blood.

2005;106(9):3160.

51. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA: a cancer journal

for clinicians. 2012;62(1):10-29.

52. Palumbo A, Bringhen S, Ludwig H, Dimopoulos MA, Bladé J, Mateos MV, et al.

Personalized therapy in multiple myeloma according to patient age and vulnerability:

a report of the European Myeloma Network (EMN). Blood. 2011;118(17):4519-29.

53. Nau K, Lewis W. Multiple myeloma: diagnosis and treatment. American family

physician. 2008;78(7):853.

References Page 267

54. Cavo M, Rajkumar SV, Palumbo A, Moreau P, Orlowski R, Bladé J, et al.

International Myeloma Working Group consensus approach to the treatment of

multiple myeloma patients who are candidates for autologous stem cell

transplantation. Blood. 2011;117(23):6063-73.

55. Mitsiades N, Mitsiades CS, Poulaki V, Chauhan D, Richardson PG, Hideshima

T, et al. Apoptotic signaling induced by immunomodulatory thalidomide analogs in

human multiple myeloma cells: therapeutic implications. Blood. 2002;99(12):4525-

30.

56. Hideshima T, Chauhan D, Shima Y, Raje N, Davies FE, Tai Y-T, et al.

Thalidomide and its analogs overcome drug resistance of human multiple myeloma

cells to conventional therapy. Blood. 2000;96(9):2943-50.

57. Zangari M, Yaccoby S, Cavallo F, Esseltine D, Tricot G. Response to bortezomib

and activation of osteoblasts in multiple myeloma. Clinical Lymphoma and Myeloma.

2006;7(2):109-14.

58. Roccaro AM, Hideshima T, Raje N, Kumar S, Ishitsuka K, Yasui H, et al.

Bortezomib mediates antiangiogenesis in multiple myeloma via direct and indirect

effects on endothelial cells. Cancer Research. 2006;66(1):184-91.

59. Yeh H, Berenson J. Myeloma bone disease and treatment options. European

Journal of Cancer. 2006;42(11):1554-63.

60. Russell R. Bisphosphonates: mode of action and pharmacology. Pediatrics.

2007;119(Supplement):S150.

61. Luckman S, Hughes D, Coxon F, Russell R, Rogers M. Nitrogen Containing

Biphosphonates Inhibit the Mevalonate Pathway and Prevent Post Translational

Prenylation of GTP Binding Proteins, Including Ras. Journal of Bone and Mineral

Research. 2005;20(7):1265-74.

References Page 268

62. . !!! INVALID CITATION !!!

63. Shipman C, Rogers M, Apperley J, Russell R, Croucher P. Bisphosphonates

induce apoptosis in human myeloma cell lines: a novel anti-tumour activity. British

journal of haematology. 1997;98(3):665-72.

64. Croucher P, De Raeve H, Perry M, Hijzen A, Shipman C, Lippitt J, et al.

Zoledronic Acid Treatment of 5T2MM Bearing Mice Inhibits the Development of

Myeloma Bone Disease: Evidence for Decreased Osteolysis, Tumor Burden and

Angiogenesis, and Increased Survival. Journal of Bone and Mineral Research.

2003;18(3):482-92.

65. Kellner J, Liu B, Kang Y, Li Z. Fact or fiction-identifying the elusive multiple

myeloma stem cell. Journal of hematology & oncology. 2013;6(1):1-10.

66. Prabhala RH, Neri P, Bae JE, Tassone P, Shammas MA, Allam CK, et al.

Dysfunctional T regulatory cells in multiple myeloma. Blood. 2006;107(1):301-4.

67. Dhodapkar MV, Geller MD, Chang DH, Shimizu K, Fujii S-I, Dhodapkar KM, et

al. A reversible defect in natural killer T cell function characterizes the progression of

premalignant to malignant multiple myeloma. The Journal of experimental medicine.

2003;197(12):1667-76.

68. Moser K, Tokoyoda K, Radbruch A, MacLennan I, Manz RA. Stromal niches,

plasma cell differentiation and survival. Current opinion in immunology.

2006;18(3):265-70.

69. Wang LD, Wagers AJ. Dynamic niches in the origination and differentiation of

haematopoietic stem cells. Nature Reviews Molecular Cell Biology. 2011;12(10):643-

55.

70. Chen Z, Orlowski RZ, Wang M, Kwak L, McCarty N. Osteoblastic niche supports

the growth of quiescent multiple myeloma cells. Blood. 2014:blood-2013-07-517136.

References Page 269

71. Visnjic D, Kalajzic Z, Rowe D, Katavic V, Lorenzo J, Aguila H. Hematopoiesis is

severely altered in mice with an induced osteoblast deficiency. Blood.

2004;103(9):3258.

72. Schofield R. The relationship between the spleen colony-forming cell and the

haemopoietic stem cell. Blood cells. 1978;4(1-2):7.

73. Nilsson S, Johnston H, Coverdale J. Spatial localization of transplanted

hemopoietic stem cells: inferences for the localization of stem cell niches. Blood.

2001;97(8):2293.

74. Reya T, Morrison S, Clarke M, Weissman I. Stem cells, cancer, and cancer stem

cells. Nature. 2001;414(6859):105-11.

75. Suda T, Arai F, Hirao A. Hematopoietic stem cells and their niche. Trends in

immunology. 2005;26(8):426-33.

76. Shiozawa Y, Taichman RS. Getting blood from bone: An emerging

understanding of the role that osteoblasts play in regulating hematopoietic stem cells

within their niche. Experimental Hematology. 2012;40(9):685-94.

77. Kiel MJ, Radice GL, Morrison SJ. Lack of evidence that hematopoietic stem

cells depend on N-cadherin-mediated adhesion to osteoblasts for their maintenance.

Cell stem cell. 2007;1(2):204-17.

78. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, et al. Targeted

disruption of Cbfa1 results in a complete lack of bone formation owing to

maturational arrest of osteoblasts. Cell. 1997;89(5):755-64.

79. Deguchi K, Yagi H, Inada M, Yoshizaki K, Kishimoto T, Komori T. Excessive

extramedullary hematopoiesis in Cbfa1-deficient mice with a congenital lack of bone

marrow. Biochemical and biophysical research communications. 1999;255(2):352-9.

References Page 270

80. Visnjic D, Kalajzic I, Gronowicz G, Aguila H, Clark S, Lichtler A, et al. Conditional

ablation of the osteoblast lineage in Col2. 3 tk transgenic mice. Journal of Bone and

Mineral Research. 2001;16(12):2222-31.

81. Calvi L, Sims N, Hunzelman J, Knight M, Giovannetti A, Saxton J, et al.

Activated parathyroid hormone/parathyroid hormone–related protein receptor in

osteoblastic cells differentially affects cortical and trabecular bone. Journal of Clinical

Investigation. 2001;107(3):277-86.

82. Calvi L, Adams G, Weibrecht K, Weber J, Olson D, Knight M, et al. Osteoblastic

cells regulate the haematopoietic stem cell niche. Nature. 2003;425(6960):841-6.

83. Zhang J, Niu C, Ye L, Huang H, He X, Tong W, et al. Identification of the

haematopoietic stem cell niche and control of the niche size. Nature.

2003;425(6960):836-41.

84. Xie Y, Yin T, Wiegraebe W, He X, Miller D, Stark D, et al. Detection of functional

haematopoietic stem cell niche using real-time imaging. Nature. 2008;457(7225):97-

101.

85. Arai F, Hirao A, Ohmura M, Sato H, Matsuoka S, Takubo K, et al.

Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the

bone marrow niche. Cell. 2004;118(2):149-61.

86. Kopp H, Avecilla S, Hooper A, Rafii S. The bone marrow vascular niche: home

of HSC differentiation and mobilization. Physiology. 2005;20(5):349.

87. Li W, Johnson S, Shelley W, Yoder M. Hematopoietic stem cell repopulating

ability can be maintained in vitro by some primary endothelial cells. Experimental

hematology. 2004;32(12):1226-37.

References Page 271

88. Cardier J, Barberá-Guillem E. Extramedullary hematopoiesis in the adult

mouse liver is associated with specific hepatic sinusoidal endothelial cells.

Hepatology. 1997;26(1):165-75.

89. Ding L, Saunders TL, Enikolopov G, Morrison SJ. Endothelial and perivascular

cells maintain haematopoietic stem cells. Nature. 2012;481(7382):457-62.

90. Bacon K, Baggiolini M, Broxmeyer H, Horuk R, Lindley I, Mantovani A, et al.

Chemokine/chemokine receptor nomenclature. Journal of Leukocyte Biology.

2001;70(3):465.

91. Baggiolini M, Dewald B, Moser B. Human chemokines: an update.

Immunology. 1997;15(1):675.

92. Kortesidis A, Zannettino A, Isenmann S, Shi S, Lapidot T, Gronthos S. Stromal-

derived factor-1 promotes the growth, survival, and development of human bone

marrow stromal stem cells. Blood. 2005;105(10):3793.

93. Dar A, Goichberg P, Shinder V, Kalinkovich A, Kollet O, Netzer N, et al.

Chemokine receptor CXCR4–dependent internalization and resecretion of functional

chemokine SDF-1 by bone marrow endothelial and stromal cells. Nature

immunology. 2005;6(10):1038-46.

94. Sugiyama T, Kohara H, Noda M, Nagasawa T. Maintenance of the

hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow

stromal cell niches. Immunity. 2006;25(6):977-88.

95. Balabanian K, Lagane B, Infantino S, Chow KY, Harriague J, Moepps B, et al.

The chemokine SDF-1/CXCL12 binds to and signals through the orphan receptor

RDC1 in T lymphocytes. Journal of Biological Chemistry. 2005;280(42):35760-6.

96. Rosenbaum DM, Rasmussen SG, Kobilka BK. The structure and function of G-

protein-coupled receptors. Nature. 2009;459(7245):356-63.

References Page 272

97. Wu B, Chien EY, Mol CD, Fenalti G, Liu W, Katritch V, et al. Structures of the

CXCR4 chemokine GPCR with small-molecule and cyclic peptide antagonists. Science.

2010;330(6007):1066-71.

98. Rosu-Myles M, Gallacher L, Murdoch B, Hess D, Keeney M, Kelvin D, et al. The

human hematopoietic stem cell compartment is heterogeneous for CXCR4

expression. Proceedings of the National Academy of Sciences of the United States of

America. 2000;97(26):14626.

99. Wright D, Bowman E, Wagers A, Butcher E, Weissman I. Hematopoietic stem

cells are uniquely selective in their migratory response to chemokines. The Journal

of experimental medicine. 2002;195(9):1145.

100. Teicher BA, Fricker SP. CXCL12 (SDF-1)/CXCR4 pathway in cancer. Clinical

Cancer Research. 2010;16(11):2927-31.

101. Thelen M. Dancing to the tune of chemokines. Nature immunology.

2001;2(2):129-34.

102. Sodhi A, Montaner S, Gutkind JS. Viral hijacking of G-protein-coupled-

receptor signalling networks. Nature Reviews Molecular Cell Biology.

2004;5(12):998-1012.

103. Bendall LJ, Baraz R, Juarez J, Shen W, Bradstock KF. Defective p38 mitogen-

activated protein kinase signaling impairs chemotaxic but not proliferative responses

to stromal-derived factor-1α in acute lymphoblastic leukemia. Cancer research.

2005;65(8):3290-8.

104. Domanska UM, Timmer-Bosscha H, Nagengast WB, Munnink THO, Kruizinga

RC, Ananias HJ, et al. CXCR4 inhibition with AMD3100 sensitizes prostate cancer to

docetaxel chemotherapy. Neoplasia (New York, NY). 2012;14(8):709.

References Page 273

105. Wang J-F, Park I-W, Groopman JE. Stromal cell-derived factor-1α stimulates

tyrosine phosphorylation of multiple focal adhesion proteins and induces migration

of hematopoietic progenitor cells: roles of phosphoinositide-3 kinase and protein

kinase C. Blood. 2000;95(8):2505-13.

106. Rossi D, Zlotnik A. The biology of chemokines and their receptors. Annual

review of immunology. 2000;18(1):217-42.

107. Yin T, Li L. The stem cell niches in bone. Journal of Clinical Investigation.

2006;116(5):1195-201.

108. Nagasawa T, Hirota S, Tachibana K, Takakura N, Nishikawa S, Kitamura Y, et

al. Defects of B-cell lymphopoiesis and bone-marrow myelopoiesis in mice lacking

the CXC chemokine PBSF/SDF-1. 1996.

109. Ara T, Tokoyoda K, Sugiyama T, Egawa T, Kawabata K, Nagasawa T. Long-term

hematopoietic stem cells require stromal cell-derived factor-1 for colonizing bone

marrow during ontogeny. Immunity. 2003;19(2):257-67.

110. Tzeng YS, Li H, Kang YL, Chen WC, Cheng WC, Lai DM. Loss of Cxcl12/Sdf-1 in

adult mice decreases the quiescent state of hematopoietic stem/progenitor cells and

alters the pattern of hematopoietic regeneration after myelosuppression. Blood.

2011;117(2):429.

111. Ponomaryov T, Peled A, Petit I, Taichman R, Habler L, Sandbank J, et al.

Induction of the chemokine stromal-derived factor-1 following DNA damage

improves human stem cell function. Journal of Clinical Investigation.

2000;106(11):1331-9.

112. Peled A, Petit I, Kollet O, Magid M, Ponomaryov T, Byk T, et al. Dependence

of human stem cell engraftment and repopulation of NOD/SCID mice on CXCR4.

Science. 1999;283(5403):845.

References Page 274

113. Walker L, Lynch M, Silverman S, Fraser J, Boulter J, Weinmaster G, et al. The

Notch/Jagged pathway inhibits proliferation of human hematopoietic progenitors in

vitro. Stem Cells. 1999;17(3):162-71.

114. Karanu F, Murdoch B, Gallacher L, Wu D, Koremoto M, Sakano S, et al. The

notch ligand jagged-1 represents a novel growth factor of human hematopoietic

stem cells. The Journal of experimental medicine. 2000;192(9):1365.

115. Stier S, Cheng T, Dombkowski D, Carlesso N, Scadden D. Notch1 activation

increases hematopoietic stem cell self-renewal in vivo and favors lymphoid over

myeloid lineage outcome. Blood. 2002;99(7):2369.

116. Artavanis-Tsakonas S, Rand M, Lake R. Notch signaling: cell fate control and

signal integration in development. Science. 1999;284(5415):770.

117. Baron M, editor An overview of the Notch signalling pathway2003: Elsevier.

118. Weinmaster G. The ins and outs of notch signaling. Molecular and Cellular

Neuroscience. 1997;9(2):91-102.

119. Hadland B, Huppert S, Kanungo J, Xue Y, Jiang R, Gridley T, et al. A

requirement for Notch1 distinguishes 2 phases of definitive hematopoiesis during

development. Blood. 2004;104(10):3097.

120. Milner L, Kopan R, Martin D, Bernstein I. A human homologue of the

Drosophila developmental gene, Notch, is expressed in CD34+ hematopoietic

precursors. Blood. 1994;83(8):2057.

121. Swiatek P, Lindsell C, del Amo F, Weinmaster G, Gridley T. Notch1 is essential

for postimplantation development in mice. Genes & development. 1994;8(6):707.

References Page 275

122. Varnum-Finney B, Purton L, Yu M, Brashem-Stein C, Flowers D, Staats S, et al.

The Notch ligand, Jagged-1, influences the development of primitive hematopoietic

precursor cells. Blood. 1998;91(11):4084.

123. Kovall RA, Blacklow SC. Chapter Two-Mechanistic Insights into Notch

Receptor Signaling from Structural and Biochemical Studies. Current topics in

developmental biology. 2010;92:31-71.

124. Burns C, Traver D, Mayhall E, Shepard J, Zon L. Hematopoietic stem cell fate

is established by the Notch–Runx pathway. Genes & development.

2005;19(19):2331.

125. Kopan R, Ilagan MXG. The canonical Notch signaling pathway: unfolding the

activation mechanism. Cell. 2009;137(2):216-33.

126. Radtke F, Wilson A, Stark G, Bauer M, van Meerwijk J, MacDonald H, et al.

Deficient T cell fate specification in mice with an induced inactivation of Notch1.

Immunity. 1999;10(5):547-58.

127. Uchida T, Nakashima M, Hirota Y, Miyazaki Y, Tsukazaki T, Shindo H.

Immunohistochemical localisation of protein tyrosine kinase receptors Tie-1 and Tie-

2 in synovial tissue of rheumatoid arthritis: correlation with angiogenesis and

synovial proliferation. Annals of the rheumatic diseases. 2000 Aug;59(8):607-14.

PubMed PMID: 10913057. Pubmed Central PMCID: PMC1753215. Epub 2000/07/27.

eng.

128. Brkovic A, Pelletier M, Girard D, Sirois MG. Angiopoietin chemotactic activities

on neutrophils are regulated by PI-3K activation. J Leukoc Biol. 2007 Apr;81(4):1093-

101. PubMed PMID: 17215522. Epub 2007/01/12. eng.

129. Kosacka J, Figiel M, Engele J, Hilbig H, Majewski M, Spanel-Borowski K.

Angiopoietin-1 promotes neurite outgrowth from dorsal root ganglion cells positive

References Page 276

for Tie-2 receptor. Cell and tissue research. 2005 Apr;320(1):11-9. PubMed PMID:

15714275. Epub 2005/02/17. eng.

130. Sacchetti B, Funari A, Michienzi S, Di Cesare S, Piersanti S, Saggio I, et al. Self-

renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic

microenvironment. Cell. 2007;131(2):324-36.

131. Currie MJ, Gunningham SP, Han C, Scott PA, Robinson BA, Harris AL, et al.

Angiopoietin-1 is inversely related to thymidine phosphorylase expression in human

breast cancer, indicating a role in vascular remodeling. Clinical cancer research : an

official journal of the American Association for Cancer Research. 2001 Apr;7(4):918-

27. PubMed PMID: 11309342. Epub 2001/04/20. eng.

132. Schubert SY, Benarroch A, Monter-Solans J, Edelman ER. Primary monocytes

regulate endothelial cell survival through secretion of angiopoietin-1 and activation

of endothelial Tie2. Arteriosclerosis, thrombosis, and vascular biology. 2011

Apr;31(4):870-5. PubMed PMID: 21273558. Pubmed Central PMCID: PMC3104028.

Epub 2011/01/29. eng.

133. Huang H, Bhat A, Woodnutt G, Lappe R. Targeting the ANGPT–TIE2 pathway

in malignancy. Nature Reviews Cancer. 2010;10(8):575-85.

134. Shewchuk L, Hassell A, Ellis B, Holmes W, Davis R, Horne E, et al. Structure of

the Tie2 RTK Domain:: Self-Inhibition by the Nucleotide Binding Loop, Activation

Loop, and C-Terminal Tail. Structure. 2000;8(11):1105-13.

135. Kim I, Moon S-O, Kim SH, Kim HJ, Koh YS, Koh GY. VEGF stimulates expression

of ICAM-1, VCAM-1 and E-selectin through nuclear factor-kappaB activation in

endothelial cells. Journal of Biological Chemistry. 2000.

136. Kim I, Moon S-O, Park SK, Chae SW, Koh GY. Angiopoietin-1 reduces VEGF-

stimulated leukocyte adhesion to endothelial cells by reducing ICAM-1, VCAM-1, and

E-selectin expression. Circulation research. 2001;89(6):477-9.

References Page 277

137. Fukuhara S, Sako K, Minami T, Noda K, Kim HZ, Kodama T, et al. Differential

function of Tie2 at cell–cell contacts and cell–substratum contacts regulated by

angiopoietin-1. Nature cell biology. 2008;10(5):513-26.

138. Kim I, Kim HG, So J-N, Kim JH, Kwak HJ, Koh GY. Angiopoietin-1 regulates

endothelial cell survival through the phosphatidylinositol 3′-kinase/Akt signal

transduction pathway. Circulation research. 2000;86(1):24-9.

139. Moss A. The angiopoietin:Tie 2 interaction: a potential target for future

therapies in human vascular disease. Cytokine & growth factor reviews. 2013

Dec;24(6):579-92. PubMed PMID: 23838360. Epub 2013/07/11. eng.

140. Puri M, Partanen J, Rossant J, Bernstein A. Interaction of the TEK and TIE

receptor tyrosine kinases during cardiovascular development. Development.

1999;126(20):4569.

141. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, Davis S, et al.

Requisite Role of Angiopoietin-1, a Ligand for the TIE2 Receptor, during Embryonic

Angiogenesis. Cell. 1996 12/27/;87(7):1171-80.

142. Arai F, Suda T. Maintenance of Quiescent Hematopoietic Stem Cells in the

Osteoblastic Niche. New York Academy Sciences Annals. 2007;1106:41-53.

143. Puri M, Bernstein A. Requirement for the TIE family of receptor tyrosine

kinases in adult but not fetal hematopoiesis. Proceedings of the National Academy

of Sciences. 2003;100(22):12753.

144. Steinberg M, McNutt P. Cadherins and their connections: adhesion junctions

have broader functions. Current opinion in cell biology. 1999;11(5):554-60.

145. Koch A, Manzur K, Shan W. Structure-based models of cadherin-mediated cell

adhesion: the evolution continues. Cellular and Molecular Life Sciences.

2004;61(15):1884-95.

References Page 278

146. Pettitt J. The cadherin superfamily. WormBook. 2005;29:1-9.

147. Angst B, Marcozzi C, Magee A. The cadherin superfamily: diversity in form and

function. Journal of cell science. 2001;114(4):629-42.

148. Radice G, Rayburn H, Matsunami H, Knudsen K, Takeichi M, Hynes R.

Developmental defects in mouse embryos lacking N-cadherin. Developmental

biology. 1997;181(1):64-78.

149. Reiss K, Maretzky T, Ludwig A, Tousseyn T, De Strooper B, Hartmann D, et al.

ADAM10 cleavage of N-cadherin and regulation of cell–cell adhesion and β-catenin

nuclear signalling. The EMBO journal. 2005;24(4):742-52.

150. Kiel MJ, Acar M, Radice GL, Morrison SJ. Hematopoietic stem cells do not

depend on N-cadherin to regulate their maintenance. Cell stem cell. 2009;4(2):170-

9.

151. Bromberg O, Frisch BJ, Weber JM, Porter RL, Civitelli R, Calvi LM. Osteoblastic

N-cadherin is not required for microenvironmental support and regulation of

hematopoietic stem and progenitor cells. Blood. 2012;120(2):303-13.

152. Cheng S, Lecanda F, Davidson M, Warlow P, Zhang S, Zhang L, et al. Human

Osteoblasts Express a Repertoire of Cadherins, Which Are Critical for BMP 2–Induced

Osteogenic Differentiation. Journal of Bone and Mineral Research. 1998;13(4):633-

44.

153. Hosokawa K, Arai F, Yoshihara H, Iwasaki H, Nakamura Y, Gomei Y, et al.

Knockdown of N-cadherin suppresses the long-term engraftment of hematopoietic

stem cells. Blood. 2010;116(4):554-63.

154. Hosokawa K, Arai F, Yoshihara H, Iwasaki H, Hembree M, Yin T, et al. Cadherin-

based adhesion is a potential target for niche manipulation to protect hematopoietic

stem cells in adult bone marrow. Cell Stem Cell. 2010;6(3):194-8.

References Page 279

155. Shiozawa Y, Pedersen EA, Havens AM, Jung Y, Mishra A, Joseph J, et al. Human

prostate cancer metastases target the hematopoietic stem cell niche to establish

footholds in mouse bone marrow. The Journal of clinical investigation.

2011;121(4):1298-312.

156. Wang N, Docherty FE, Brown HK, Reeves KJ, Fowles A, Ottewell PD, et al.

Prostate Cancer Cells Preferentially Home to Osteoblast‐Rich Areas in the Early

Stages of Bone Metastasis–Evidence from In Vivo Models. Journal of Bone and

Mineral Research. 2014.

157. Cattan AR, Douglas E. The CB 17 scid mouse strain as a model for

human disseminated leukaemia and myeloma in vivo. Leukemia research.

1994;18(7):513-22.

158. Bellamy WT, Mendibles P, Bontje P, Thompson F, Richter L, Weinstein RS, et

al. Development of an orthotopic SCID mouse-human tumor xenograft model

displaying the multidrug-resistant phenotype. Cancer chemotherapy and

pharmacology. 1996;37(4):305-16.

159. Huang Y-W, Richardson JA, Vitetta ES. Anti-CD54 (ICAM-1) has antitumor

activity in SCID mice with human myeloma cells. Cancer research. 1995;55(3):610-6.

160. Alsina M, Boyce B, Devlin RD, Anderson JL, Craig F, Mundy GR, et al.

Development of an in vivo model of human multiple myeloma bone disease. Blood.

1996;87(4):1495-501.

161. Bellamy WT, Odeleye A, Finley P, Huizenga B, Dalton WS, Weinstein RS, et al.

An in vivo model of human multidrug-resistant multiple myeloma in SCID mice. The

American journal of pathology. 1993;142(3):691.

162. Suzuki H, Yasukawa K, Saito T, Goitsuka R, Hasegawa A, Ohsugi Y, et al. Anti‐

human interleukin‐6 receptor antibody inhibits human myeloma growth in vivo.

European journal of immunology. 1992;22(8):1989-93.

References Page 280

163. Namikawa R, Ueda R, Kyoizumi S. Growth of human myeloid leukemias in the

human marrow environment of SCID-hu mice. Blood. 1993;82(8):2526-36.

164. Yaccoby S, Barlogie B, Epstein J. Primary myeloma cells growing in SCID-hu

mice: a model for studying the biology and treatment of myeloma and its

manifestations. Blood. 1998;92(8):2908-13.

165. Miyakawa Y, Ohnishi Y, Tomisawa M, Monnai M, Kohmura K, Ueyama Y, et al.

Establishment of a new model of human multiple myeloma using NOD/SCID/γ

c null(NOG) mice. Biochemical and biophysical research

communications. 2004;313(2):258-62.

166. Mirandola L, Yu Y, Jenkins M, Chiaramonte R, Cobos E, John C, et al. Tracking

human multiple myeloma xenografts in NOD-Rag-1/IL-2 receptor gamma chain-null

mice with the novel biomarker AKAP-4. BMC cancer. 2011;11(1):394.

167. Radl J, Croese J, Zurcher C, Van den Enden-Vieveen M, De Leeuw A. Animal

model of human disease. Multiple myeloma. The American journal of pathology.

1988;132(3):593.

168. Radl J, De Glopper ED, Schuit HR, Zurcher C. Idiopathic paraproteinemia. II.

Transplantation of the paraprotein-producing clone from old to young

C57BL/KaLwRij mice. Journal of immunology (Baltimore, Md : 1950). 1979

Feb;122(2):609-13. PubMed PMID: 368243. Epub 1979/02/01. eng.

169. Buckle CH, De Leenheer E, Lawson MA, Yong K, Rabin N, Perry M, et al. Soluble

rank ligand produced by myeloma cells causes generalised bone loss in multiple

myeloma. PloS one. 2012;7(8):e41127.

170. Menu E, Asosingh K, Indraccolo S, De Raeve H, Van Riet I, Van Valckenborgh

E, et al. The involvement of stromal derived factor 1alpha in homing and progression

of multiple myeloma in the 5TMM model. Haematologica. 2006;91(5):605.

References Page 281

171. Vanderkerken K, De Raeve H, Goes E, Van Meirvenne S, Radl J, Van Riet I, et

al. Organ involvement and phenotypic adhesion profile of 5T2 and 5T33 myeloma

cells in the C57BL/KaLwRij mouse. British journal of cancer. 1997;76(4):451.

172. Glick AF, Song YS, Hwang B, Lillvis J, Zanzonico P, Fuchs C, et al. Age matters:

Young T lymphocytes offer better protection from myeloma proliferation. Immunity

& Ageing. 2013;10(1):5.

173. Nur H, Fostier K, Aspeslagh S, Renmans W, Bertrand E, Leleu X, et al.

Preclinical Evaluation of Invariant Natural Killer T Cells in the 5T33 Multiple Myeloma

Model. PloS one. 2013;8(5):e65075.

174. Garrett I, Dallas S, Radl J, Mundy G. A murine model of human myeloma bone

disease. Bone. 1997;20(6):515-20.

175. Fowler JA, Mundy GR, Lwin ST, Edwards CM. Bone marrow stromal cells

create a permissive microenvironment for myeloma development: a new stromal

role for Wnt inhibitor Dkk1. Cancer research. 2012;72(9):2183-9.

176. Edwards CM, Lwin ST, Fowler JA, Oyajobi BO, Zhuang J, Bates AL, et al.

Myeloma cells exhibit an increase in proteasome activity and an enhanced response

to proteasome inhibition in the bone marrow microenvironment in vivo. American

journal of hematology. 2009;84(5):268-72.

177. Dairaghi DJ, Oyajobi BO, Gupta A, McCluskey B, Miao S, Powers JP, et al. CCR1

blockade reduces tumor burden and osteolysis in vivo in a mouse model of myeloma

bone disease. Blood. 2012;120(7):1449-57.

178. Noll JE, Williams SA, Tong CM, Wang H, Quach JM, Purton LE, et al. Myeloma

plasma cells alter the bone marrow microenvironment by stimulating the

proliferation of mesenchymal stromal cells. Haematologica. 2013:haematol.

2013.090977.

References Page 282

179. Asosingh K, Radl J, Van Riet I, Van Camp B, Vanderkerken K. The 5TMM series:

a useful in vivo mouse model of human multiple myeloma. The Hematology Journal.

2000;1(5):351-6.

180. Kodama H. Establishment of a clonal osteogenic cell line from newborn

mouse calvaria. Jpn J Oral Biol. 1981;23:899-901.

181. Kitching R, Qi S, Li V, Raouf A, Vary CP, Seth A. Coordinate gene expression

patterns during osteoblast maturation and retinoic acid treatment of MC3T3-E1 cells.

Journal of bone and mineral metabolism. 2002;20(5):269-80.

182. Bakker A, Klein-Nulend J. Osteoblast isolation from murine calvariae and long

bones. Bone Research Protocols: Springer; 2003. p. 19-28.

183. Lian JB, Stein GS. Development of the osteoblast phenotype: molecular

mechanisms mediating osteoblast growth and differentiation. The Iowa orthopaedic

journal. 1995;15:118.

184. Yang D, Atkins GJ, Turner AG, Anderson PH, Morris HA. Differential effects of

1, 25-dihydroxyvitamin D on mineralisation and differentiation in two different types

of osteoblast-like cultures. The Journal of steroid biochemistry and molecular

biology. 2013;136:166-70.

185. Brown M, Wittwer C. Flow cytometry: principles and clinical applications in

hematology. Clinical chemistry. 2000;46(8):1221-9.

186. Sabokbar A, Millett P, Myer B, Rushton N. A rapid, quantitative assay for

measuring alkaline phosphatase activity in osteoblastic cells in vitro. Bone and

mineral. 1994;27(1):57-67.

187. Wang N, Rumney RM, Yang L, Robaye B, Boeynaems JM, Skerry TM, et al. The

P2Y13 receptor regulates extracellular ATP metabolism and the osteogenic response

to mechanical loading. Journal of Bone and Mineral Research. 2013;28(6):1446-56.

References Page 283

188. Singer O, Verma IM. Applications of lentiviral vectors for shRNA delivery and

transgenesis. Current gene therapy. 2008;8(6):483.

189. Manjunath N, Wu H, Subramanya S, Shankar P. Lentiviral delivery of short

hairpin RNAs. Advanced drug delivery reviews. 2009;61(9):732-45.

190. MacRae IJ, Zhou K, Li F, Repic A, Brooks AN, Cande WZ, et al. Structural basis

for double-stranded RNA processing by Dicer. Science. 2006;311(5758):195-8.

191. Wang Z, Rao DD, Senzer N, Nemunaitis J. RNA interference and cancer

therapy. Pharmaceutical research. 2011;28(12):2983-95.

192. Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Müller R.

Guidelines for assessment of bone microstructure in rodents using micro–computed

tomography. Journal of Bone and Mineral Research. 2010;25(7):1468-86.

193. Nakayama T, Hieshima K, Izawa D, Tatsumi Y, Kanamaru A, Yoshie O. Cutting

edge: profile of chemokine receptor expression on human plasma cells accounts for

their efficient recruitment to target tissues. The Journal of Immunology.

2003;170(3):1136.

194. Möller C, Strömberg T, Juremalm M, Nilsson K, Nilsson G. Expression and

function of chemokine receptors in human multiple myeloma. Leukemia.

2003;17(1):203-10.

195. Alsayed Y, Ngo H, Runnels J, Leleu X, Singha U, Pitsillides C, et al. Mechanisms

of regulation of CXCR4/SDF-1 (CXCL12)-dependent migration and homing in multiple

myeloma. Blood. 2007;109(7):2708.

196. Jung Y, Wang J, Schneider A, Sun Y-X, Koh-Paige A, Osman N, et al. Regulation

of SDF-1 (CXCL12) production by osteoblasts; a possible mechanism for stem cell

homing. Bone. 2006;38(4):497-508.

References Page 284

197. Nefedova Y, Cheng P, Alsina M, Dalton WS, Gabrilovich DI. Involvement of

Notch-1 signaling in bone marrow stroma-mediated de novo drug resistance of

myeloma and other malignant lymphoid cell lines. Blood. 2004;103(9):3503-10.

198. Jundt F, Probsting K, Anagnostopoulos I, Muehlinghaus G, Chatterjee M,

Mathas S, et al. Jagged1-induced Notch signaling drives proliferation of multiple

myeloma cells. Blood. 2004;103(9):3511.

199. Nefedova Y, Cheng P, Alsina M, Dalton W, Gabrilovich D. Involvement of

Notch-1 signaling in bone marrow stroma-mediated de novo drug resistance of

myeloma and other malignant lymphoid cell lines. Blood. 2004;103(9):3503.

200. Pereira RMR, Delany AM, Durant D, Canalis E. Cortisol regulates the

expression of Notch in osteoblasts. Journal of cellular biochemistry. 2002;85(2):252-

8.

201. Giuliani N, Colla S, Lazzaretti M, Sala R, Roti G, Mancini C, et al. Proangiogenic

properties of human myeloma cells: production of angiopoietin-1 and its potential

relationship to myeloma-induced angiogenesis. Blood. 2003;102(2):638.

202. Jeong B-C, Kim H-J, Bae I-H, Lee K-N, Lee K-Y, Oh W-M, et al. COMP-Ang1, a

chimeric form of Angiopoietin 1, enhances BMP2-induced osteoblast differentiation

and bone formation. Bone. 2010;46(2):479-86.

203. Tanaka H, Kono E, Tran CP, Miyazaki H, Yamashiro J, Shimomura T, et al.

Monoclonal antibody targeting of N-cadherin inhibits prostate cancer growth,

metastasis and castration resistance. Nature medicine. 2010;16(12):1414-20.

204. Nieman M, Prudoff R, Johnson K, Wheelock M. N-cadherin promotes motility

in human breast cancer cells regardless of their E-cadherin expression. The Journal

of cell biology. 1999;147(3):631.

References Page 285

205. Vandyke K, Chow AW, Williams SA, To LB, Zannettino AC. Circulating N‐

cadherin levels are a negative prognostic indicator in patients with multiple

myeloma. British Journal of Haematology. 2013.

206. Groen RW, de Rooij MF, Kocemba KA, Reijmers RM, de Haan-Kramer A,

Overdijk MB, et al. N-cadherin-mediated interaction with multiple myeloma cells

inhibits osteoblast differentiation. Haematologica. 2011;96(11):1653-61.

207. Sadler NM, Harris BR, Metzger BA, Kirshner J. N-cadherin impedes

proliferation of the multiple myeloma cancer stem cells. American journal of blood

research. 2013;3(4):271.

208. Tsutsumimoto T, Kawasaki S, Ebara S, Takaoka K. TNF and IL 1 Suppress N

Cadherin Expression in MC3T3 E1 Cells. Journal of Bone and Mineral Research.

1999;14(10):1751-60.

209. Vogel C, Marcotte EM. Insights into the regulation of protein abundance from

proteomic and transcriptomic analyses. Nature Reviews Genetics. 2012;13(4):227-

32.

210. Maier T, Güell M, Serrano L. Correlation of mRNA and protein in complex

biological samples. FEBS letters. 2009;583(24):3966-73.

211. Schwanhäusser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, et al. Global

quantification of mammalian gene expression control. Nature. 2011;473(7347):337-

42.

212. Greenbaum D, Colangelo C, Williams K, Gerstein M. Comparing protein

abundance and mRNA expression levels on a genomic scale. Genome Biol.

2003;4(9):117.

References Page 286

213. Hatse S, Balzarini J, Liekens S. Stromal cell-derived factor 1 (CXCL12) binds to

endothelial cells and signals through a receptor different from CXCR4. Biochemical

and biophysical research communications. 2006;348(1):192-9.

214. Skrtic A, Korac P, Kristo DR, Ajdukovic Stojisavljevic R, Ivankovic D, Dominis M.

Immunohistochemical analysis of NOTCH1 and JAGGED1 expression in multiple

myeloma and monoclonal gammopathy of undetermined significance. Human

pathology. 2010.

215. Zannettino AC, Farrugia AN, Kortesidis A, Manavis J, To LB, Martin SK, et al.

Elevated serum levels of stromal-derived factor-1α are associated with increased

osteoclast activity and osteolytic bone disease in multiple myeloma patients. Cancer

research. 2005;65(5):1700-9.

216. Nobta M, Tsukazaki T, Shibata Y, Xin C, Moriishi T, Sakano S, et al. Critical

regulation of bone morphogenetic protein-induced osteoblastic differentiation by

Delta1/Jagged1-activated Notch1 signaling. Journal of Biological Chemistry.

2005;280(16):15842-8.

217. Greenbaum A, Hsu Y-MS, Day RB, Schuettpelz LG, Christopher MJ, Borgerding

JN, et al. CXCL12 in early mesenchymal progenitors is required for haematopoietic

stem-cell maintenance. Nature. 2013;495(7440):227-30.

218. Kawaguchi J, Kii I, Sugiyama Y, Takeshita S, Kudo A. The Transition of Cadherin

Expression in Osteoblast Differentiation from Mesenchymal Cells: Consistent

Expression of Cadherin 11 in Osteoblast Lineage. Journal of Bone and Mineral

Research. 2001;16(2):260-9.

219. Ferrari SL, Traianedes K, Thorne M, Lafage‐Proust MH, Genever P, Cecchini

MG, et al. A Role for N‐Cadherin in the Development of the Differentiated

Osteoblastic Phenotype. Journal of Bone and Mineral Research. 2000;15(2):198-208.

References Page 287

220. Van de Broek I, Leleu X, Schots R, Facon T, Vanderkerken K, Van Camp B, et

al. Clinical significance of chemokine receptor (CCR1, CCR2 and CXCR4) expression in

human myeloma cells: the association with disease activity and survival.

Haematologica. 2006;91(2):200.

221. Pellegrino A, Ria R, Pietro GD, Cirulli T, Surico G, Pennisi A, et al. Bone marrow

endothelial cells in multiple myeloma secrete CXC‐chemokines that mediate

interactions with plasma cells. British journal of haematology. 2005;129(2):248-56.

222. Dallas SL, Garrett IR, Oyajobi BO, Dallas MR, Boyce BF, Bauss F, et al.

Ibandronate reduces osteolytic lesions but not tumor burden in a murine model of

myeloma bone disease. Blood. 1999;93(5):1697-706.

223. Oyajobi BO, Muñoz S, Kakonen R, Williams PJ, Gupta A, Wideman CL, et al.

Detection of myeloma in skeleton of mice by whole-body optical fluorescence

imaging. Molecular cancer therapeutics. 2007;6(6):1701-8.

224. Liu Z-J, Xiao M, Balint K, Smalley KS, Brafford P, Qiu R, et al. Notch1 signaling

promotes primary melanoma progression by activating mitogen-activated protein

kinase/phosphatidylinositol 3-kinase-Akt pathways and up-regulating N-cadherin

expression. Cancer research. 2006;66(8):4182-90.

225. Zhu H, Zhou X, Redfield S, Lewin J, Miele L. Elevated Jagged-1 and Notch-1

expression in high grade and metastatic prostate cancers. American journal of

translational research. 2013;5(3):368.

226. Santagata S, Demichelis F, Riva A, Varambally S, Hofer MD, Kutok JL, et al.

JAGGED1 expression is associated with prostate cancer metastasis and recurrence.

Cancer research. 2004;64(19):6854-7.

227. Naumann U, Cameroni E, Pruenster M, Mahabaleshwar H, Raz E, Zerwes H-G,

et al. CXCR7 functions as a scavenger for CXCL12 and CXCL11. PloS one.

2010;5(2):e9175.

References Page 288

228. Jennbacken K, Tešan T, Wang W, Gustavsson H, Damber J-E, Welén K. N-

cadherin increases after androgen deprivation and is associated with metastasis in

prostate cancer. Endocrine-Related Cancer. 2010;17(2):469-79.

229. Speetjens FM, Liefers GJ, Korbee CJ, Mesker WE, van de Velde CJ, van

Vlierberghe RL, et al. Nuclear localization of CXCR4 determines prognosis for

colorectal cancer patients. Cancer Microenvironment. 2009;2(1):1-7.

230. Yao X, Zhou L, Han S, Chen Y. High expression of CXCR4 and CXCR7 predicts

poor survival in gallbladder cancer. Journal of International Medical Research.

2011;39(4):1253-64.

231. Eaton CL, Colombel M, van der Pluijm G, Cecchini M, Wetterwald A, Lippitt J,

et al. Evaluation of the frequency of putative prostate cancer stem cells in primary

and metastatic prostate cancer. The Prostate. 2010;70(8):875-82.

232. Clarke B. Normal bone anatomy and physiology. Clinical journal of the

American Society of Nephrology. 2008;3(Supplement 3):S131-S9.

233. Friedenstein A, Chailakhyan R, Gerasimov U. Bone marrow osteogenic stem

cells: in vitro cultivation and transplantation in diffusion chambers. Cell Proliferation.

1987;20(3):263-72.

234. Harada S-i, Rodan GA. Control of osteoblast function and regulation of bone

mass. Nature. 2003;423(6937):349-55.

235. Caplan AI, Bruder SP. Mesenchymal stem cells: building blocks for molecular

medicine in the 21st century. Trends in molecular medicine. 2001;7(6):259-64.

236. Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, et al. 

Cbfa1, a Candidate Gene for Cleidocranial Dysplasia Syndrome, Is Essential for

Osteoblast Differentiation and Bone Development. Cell. 1997;89(5):765-71.

References Page 289

237. Nakamura H. Morphology, function, and differentiation of bone cells. Journal

of hard tissue biology. 2007;16(1):15-22.

238. Nakashima K, Zhou X, Kunkel G, Zhang Z, Deng JM, Behringer RR, et al. The

novel zinc finger-containing transcription factor osterix is required for osteoblast

differentiation and bone formation. Cell. 2002;108(1):17-29.

239. Hoemann C, El-Gabalawy H, McKee M. In vitro osteogenesis assays: influence

of the primary cell source on alkaline phosphatase activity and mineralization.

Pathologie Biologie. 2009;57(4):318-23.

240. Stein GS, Lian JB, Owen TA. Relationship of cell growth to the regulation of

tissue-specific gene expression during osteoblast differentiation. The FASEB journal.

1990;4(13):3111-23.

241. Choi K-M, Seo Y-K, Yoon H-H, Song K-Y, Kwon S-Y, Lee H-S, et al. Effect of

ascorbic acid on bone marrow-derived mesenchymal stem cell proliferation and

differentiation. Journal of bioscience and bioengineering. 2008;105(6):586-94.

242. Jaiswal N, Haynesworth SE, Caplan AI, Bruder SP. Osteogenic differentiation

of purified, culture‐expanded human mesenchymal stem cells in vitro. Journal of

cellular biochemistry. 1997;64(2):295-312.

243. Ecarot-Charrier B, Glorieux FH, Van Der Rest M, Pereira G. Osteoblasts

isolated from mouse calvaria initiate matrix mineralization in culture. The Journal of

cell biology. 1983;96(3):639-43.

244. Bellows C, Aubin J, Heersche J. Initiation and progression of mineralization of

bone nodules formed in vitro: the role of alkaline phosphatase and organic

phosphate. Bone and mineral. 1991;14(1):27-40.

245. Wang D, Christensen K, Chawla K, Xiao G, Krebsbach PH, Franceschi RT.

Isolation and Characterization of MC3T3‐E1 Preosteoblast Subclones with Distinct In

References Page 290

Vitro and In Vivo Differentiation/Mineralization Potential. Journal of Bone and

Mineral Research. 1999;14(6):893-903.

246. Otsuka E, Yamaguchi A, Hirose S, Hagiwara H. Characterization of osteoblastic

differentiation of stromal cell line ST2 that is induced by ascorbic acid. American

Journal of Physiology-Cell Physiology. 1999;277(1):C132-C8.

247. Gronthos S, Graves S, Ohta S, Simmons P. The STRO-1+ fraction of adult

human bone marrow contains the osteogenic precursors. Blood. 1994;84(12):4164-

73.

248. Igarashi M, Kamiya N, Hasegawa M, Kasuya T, Takahashi T, Takagi M.

Inductive effects of dexamethasone on the gene expression of Cbfa1, Osterix and

bone matrix proteins during differentiation of cultured primary rat osteoblasts.

Journal of molecular histology. 2004;35(1):3-10.

249. Peter SJ, Liang CR, Kim DJ, Widmer MS, Mikos AG. Osteoblastic phenotype of

rat marrow stromal cells cultured in the presence of dexamethasone, β‐

glycerolphosphate, and L‐ascorbic acid. Journal of cellular biochemistry.

1998;71(1):55-62.

250. Lynch MP, Stein JL, Stein GS, Lian JB. The influence of type I collagen on the

development and maintenance of the osteoblast phenotype in primary and passaged

rat calvarial osteoblasts: modification of expression of genes supporting cell growth,

adhesion, and extracellular matrix mineralization. Experimental cell research.

1995;216(1):35-45.

251. Choi JY, Lee BH, Song KB, Park RW, Kim IS, Sohn KY, et al. Expression patterns

of bone‐related proteins during osteoblastic differentiation in MC3T3‐E1 cells.

Journal of cellular biochemistry. 1996;61(4):609-18.

252. Gartland A, Rumney RM, Dillon JP, Gallagher JA. Isolation and Culture of

Human Osteoblasts. Human Cell Culture Protocols: Springer; 2012. p. 337-55.

References Page 291

253. Coelho M, Fernandes M. Human bone cell cultures in biocompatibility testing.

Part II: effect of ascorbic acid, β-glycerophosphate and dexamethasone on

osteoblastic differentiation. Biomaterials. 2000;21(11):1095-102.

254. Nefussi J-R, Boy-Lefevre M, Boulekbache H, Forest N. Mineralization in vitro

of matrix formed by osteoblasts isolated by collagenase digestion. Differentiation.

1985;29(2):160-8.

255. Shui C, Spelsberg TC, Riggs BL, Khosla S. Changes in Runx2/Cbfa1 expression

and activity during osteoblastic differentiation of human bone marrow stromal cells.

Journal of Bone and Mineral Research. 2003;18(2):213-21.

256. Sudhakar S, Li Y, Katz MS, Elango N. Translational regulation is a control point

in RUNX2/Cbfa1 gene expression. Biochemical and biophysical research


257. PAUTKE C, SCHIEKER M, TISCHER T, KOLK A, NETH P, MUTSCHLER W, et al.

Characterization of osteosarcoma cell lines MG-63, Saos-2 and U-2 OS in comparison

to human osteoblasts. Anticancer research. 2004;24(6):3743-8.

258. Taichman RS, Cooper C, Keller ET, Pienta KJ, Taichman NS, McCauley LK. Use

of the stromal cell-derived factor-1/CXCR4 pathway in prostate cancer metastasis to

bone. Cancer Research. 2002;62(6):1832-7.

259. Weber JM, Forsythe SR, Christianson CA, Frisch BJ, Gigliotti BJ, Jordan CT, et

al. Parathyroid hormone stimulates expression of the Notch ligand Jagged1 in

osteoblastic cells. Bone. 2006;39(3):485-93.

260. Greenbaum AM, Revollo LD, Woloszynek JR, Civitelli R, Link DC. N-cadherin in

osteolineage cells is not required for maintenance of hematopoietic stem cells.

Blood. 2012;120(2):295-302.

References Page 292

261. Jung Y, Wang J, Song J, Shiozawa Y, Wang J, Havens A, et al. Annexin II

expressed by osteoblasts and endothelial cells regulates stem cell adhesion, homing,

and engraftment following transplantation. Blood. 2007;110(1):82-90.

262. Barille S, Collette M, Bataille R, Amiot M. Myeloma cells upregulate

interleukin-6 secretion in osteoblastic cells through cell-to-cell contact but

downregulate osteocalcin. Blood. 1995;86(8):3151-9.

263. Fulciniti M, Tassone P, Hideshima T, Vallet S, Nanjappa P, Ettenberg SA, et al.

Anti-DKK1 mAb (BHQ880) as a potential therapeutic agent for multiple myeloma.

Blood. 2009;114(2):371-9.

264. Singh S, Singh UP, Grizzle WE, Lillard JW. CXCL12–CXCR4 interactions

modulate prostate cancer cell migration, metalloproteinase expression and invasion.

Laboratory investigation. 2004;84(12):1666-76.

265. Chen Y, Stamatoyannopoulos G, Song C-Z. Down-regulation of CXCR4 by

inducible small interfering RNA inhibits breast cancer cell invasion in vitro. Cancer

research. 2003;63(16):4801-4.

266. Fernandis AZ, Prasad A, Band H, Klösel R, Ganju RK. Regulation of CXCR4-

mediated chemotaxis and chemoinvasion of breast cancer cells. Oncogene.

2004;23(1):157-67.

267. Nefedova Y, Sullivan D, Bolick S, Dalton W, Gabrilovich D. Inhibition of Notch

signaling induces apoptosis of myeloma cells and enhances sensitivity to

chemotherapy. Blood. 2008;111(4):2220.

268. Hazan R, Phillips G, Qiao R, Norton L, Aaronson S. Exogenous expression of N-

cadherin in breast cancer cells induces cell migration, invasion, and metastasis. The

Journal of cell biology. 2000;148(4):779.

References Page 293

269. Qi J, Chen N, Wang J, Siu C-H. Transendothelial migration of melanoma cells

involves N-cadherin-mediated adhesion and activation of the β-catenin signaling

pathway. Molecular biology of the cell. 2005;16(9):4386-97.

270. Li G, Satyamoorthy K, Herlyn M. N-cadherin-mediated intercellular

interactions promote survival and migration of melanoma cells. Cancer research.

2001;61(9):3819-25.

271. Vacca A, Ribatti D, Roncali L, Ranieri G, Serio G, Silvestris F, et al. Bone marrow

angiogenesis and progression in multiple myeloma. British journal of haematology.

1994;87(3):503-8.

272. Tanaka Y, Morimoto I, Nakano Y, Okada Y, Hirota S, Nomura S, et al.

Osteoblasts are regulated by the cellular adhesion through ICAM‐1 and VCAM‐1.

Journal of Bone and Mineral Research. 1995;10(10):1462-9.

273. Kawano MM, Huang N, Tanaka H, Ishikawa H, Sakai A, Tanabe O, et al.

Homotypic cell aggregations of human myeloma cells with ICAM‐1 and LFA‐1

molecules. British journal of haematology. 1991;79(4):583-8.

274. Michigami T, Shimizu N, Williams PJ, Niewolna M, Dallas SL, Mundy GR, et al.

Cell–cell contact between marrow stromal cells and myeloma cells via VCAM-1 and

α4β1-integrin enhances production of osteoclast-stimulating activity. Blood.

2000;96(5):1953-60.

275. Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell adhesion

mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in

human myeloma cell lines. Blood. 1999;93(5):1658-67.

276. Kyle R, Durie B, Rajkumar S, Landgren O, Blade J, Merlini G, et al. Monoclonal

gammopathy of undetermined significance (MGUS) and smoldering (asymptomatic)

multiple myeloma: IMWG consensus perspectives risk factors for progression and

guidelines for monitoring and management. Leukemia. 2010;24(6):1121-7.

References Page 294

277. Abkowitz J, Robinson A, Kale S, Long M, Chen J. Mobilization of hematopoietic

stem cells during homeostasis and after cytokine exposure. Blood. 2003;102(4):1249.

278. Augustine CK, Yoshimoto Y, Gupta M, Zipfel PA, Selim MA, Febbo P, et al.

Targeting N-cadherin enhances antitumor activity of cytotoxic therapies in

melanoma treatment. Cancer research. 2008;68(10):3777-84.

279. Gravdal K, Halvorsen OJ, Haukaas SA, Akslen LA. A switch from E-cadherin to

N-cadherin expression indicates epithelial to mesenchymal transition and is of strong

and independent importance for the progress of prostate cancer. Clinical Cancer

Research. 2007;13(23):7003-11.

280. Tomita K, van Bokhoven A, van Leenders GJ, Ruijter ET, Jansen CF,

Bussemakers MJ, et al. Cadherin switching in human prostate cancer progression.

Cancer research. 2000;60(13):3650-4.

281. Vesuna F, van Diest P, Chen JH, Raman V. Twist is a transcriptional repressor

of E-cadherin gene expression in breast cancer. Biochemical and biophysical research


282. Liles WC, Broxmeyer HE, Rodger E, Wood B, Hübel K, Cooper S, et al.

Mobilization of hematopoietic progenitor cells in healthy volunteers by AMD3100, a

CXCR4 antagonist. Blood. 2003;102(8):2728-30.

283. DiPersio JF, Stadtmauer EA, Nademanee A, Micallef IN, Stiff PJ, Kaufman JL,

et al. Plerixafor and G-CSF versus placebo and G-CSF to mobilize hematopoietic stem

cells for autologous stem cell transplantation in patients with multiple myeloma.

Blood. 2009;113(23):5720-6.

284. Kuhne MR, Mulvey T, Belanger B, Chen S, Pan C, Chong C, et al. BMS-

936564/MDX-1338: A Fully Human Anti-CXCR4 Antibody Induces Apoptosis In Vitro

and Shows Antitumor Activity In Vivo in Hematologic Malignancies. Clinical Cancer

Research. 2013;19(2):357-66.

References Page 295

285. Fahham D, Weiss ID, Abraham M, Beider K, Hanna W, Shlomai Z, et al. In vitro

and in vivo therapeutic efficacy of CXCR4 antagonist BKT140 against human non–

small cell lung cancer. The Journal of Thoracic and Cardiovascular Surgery. 2012.

286. Beider K, Begin M, Abraham M, Wald H, Weiss ID, Wald O, et al. CXCR4

antagonist 4F-benzoyl-TN14003 inhibits leukemia and multiple myeloma tumor

growth. Exp Hematol. 2011 Mar;39(3):282-92. PubMed PMID: 21138752. Epub

2010/12/09. eng.

287. Wei P, Walls M, Qiu M, Ding R, Denlinger RH, Wong A, et al. Evaluation of

selective γ-secretase inhibitor PF-03084014 for its antitumor efficacy and

gastrointestinal safety to guide optimal clinical trial design. Molecular cancer

therapeutics. 2010;9(6):1618-28.

288. Samon JB, Castillo-Martin M, Hadler M, Ambesi-Impiobato A, Paietta E,

Racevskis J, et al. Preclinical analysis of the γ-secretase inhibitor PF-03084014 in

combination with glucocorticoids in T-cell acute lymphoblastic leukemia. Molecular

Cancer Therapeutics. 2012;11(7):1565-75.

289. Zhang CC, Pavlicek A, Zhang Q, Lira ME, Painter CL, Yan Z, et al. Biomarker and

Pharmacologic Evaluation of the γ-Secretase Inhibitor PF-03084014 in Breast Cancer

Models. Clinical Cancer Research. 2012;18(18):5008-19.

290. Luistro L, He W, Smith M, Packman K, Vilenchik M, Carvajal D, et al. Preclinical

profile of a potent γ-secretase inhibitor targeting notch signaling with in vivo efficacy

and pharmacodynamic properties. Cancer research. 2009;69(19):7672-80.

291. Tolcher AW, Messersmith WA, Mikulski SM, Papadopoulos KP, Kwak EL,

Gibbon DG, et al. Phase I study of RO4929097, a gamma secretase inhibitor of Notch

signaling, in patients with refractory metastatic or locally advanced solid tumors.

Journal of Clinical Oncology. 2012;30(19):2348-53.

References Page 296

292. Arcari JT, Beebe JS, Berliner MA, Bernardo V, Boehm M, Borzillo GV, et al.

Discovery and synthesis of novel 4-aminopyrrolopyrimidine Tie-2 kinase inhibitors for

the treatment of solid tumors. Bioorganic & medicinal chemistry letters. 2013.

293. Peeters M, Strickland A, Lichinitser M, Suresh A, Manikhas G, Shapiro J, et al.

A randomised, double-blind, placebo-controlled phase 2 study of trebananib (AMG

386) in combination with FOLFIRI in patients with previously treated metastatic

colorectal carcinoma. British journal of cancer. 2013.

294. Hudkins RL, Becknell NC, Zulli AL, Underiner TL, Angeles TS, Aimone LD, et al.

Synthesis and Biological Profile of the pan-Vascular Endothelial Growth Factor

Receptor/Tyrosine Kinase with Immunoglobulin and Epidermal Growth Factor-Like

Homology Domains 2 (VEGF-R/TIE-2) Inhibitor 11-(2-Methylpropyl)-12, 13-dihydro-

2-methyl-8-(pyrimidin-2-ylamino)-4 H-indazolo [5, 4-a] pyrrolo [3, 4-c] carbazol-4-

one (CEP-11981): A Novel Oncology Therapeutic Agent. Journal of medicinal

chemistry. 2012;55(2):903-13.

295. Bartee E, Chan WM, Moreb JS, Cogle CR, McFadden G. Selective purging of

human multiple myeloma cells from autologous stem cell transplantation grafts using

oncolytic myxoma virus. Biology of Blood and Marrow Transplantation.

2012;18(10):1540-51.

296. Burden-Gulley SM, Gates TJ, Craig SE, Lou SF, Oblander SA, Howell S, et al.

Novel peptide mimetic small molecules of the HAV motif in N-cadherin inhibit N-

cadherin-mediated neurite outgrowth and cell adhesion. Peptides.

2009;30(12):2380-7.

defining key molecules in a myeloma cell nicheetheses.whiterose.ac.uk/6732/1/julia phd thesis...

Documents